diff --git a/README_cn.md b/README_cn.md
index b1e41647b0efb84a1d36ff361547f6580f370202..dce469e18a5cf96d2b1e8b04b9ad3d0c85be9387 100644
--- a/README_cn.md
+++ b/README_cn.md
@@ -27,7 +27,7 @@
 
   - 发布高精度云边一体SOTA目标检测模型[PP-YOLOE](configs/ppyoloe)，提供s/m/l/x版本，l版本COCO test2017数据集精度51.6%，V100预测速度78.1 FPS，支持混合精度训练，训练较PP-YOLOv2加速33%，全系列多尺度模型，满足不同硬件算力需求，可适配服务器、边缘端GPU及其他服务器端AI加速卡。
   - 发布边缘端和CPU端超轻量SOTA目标检测模型[PP-PicoDet增强版](configs/picodet)，精度提升2%左右，CPU预测速度提升63%，新增参数量0.7M的PicoDet-XS模型，提供模型稀疏化和量化功能，便于模型加速，各类硬件无需单独开发后处理模块，降低部署门槛。
-  - 发布实时行人分析工具[PP-Human](deploy/pphuman)，支持行人跟踪、人流量统计、人体属性识别与摔倒检测四大能力，基于真实场景数据特殊优化，精准识别各类摔倒姿势，适应不同环境背景、光线及摄像角度。
+  - 发布实时行人分析工具[PP-Human](deploy/pipeline)，支持行人跟踪、人流量统计、人体属性识别与摔倒检测四大能力，基于真实场景数据特殊优化，精准识别各类摔倒姿势，适应不同环境背景、光线及摄像角度。
   - 新增[YOLOX](configs/yolox)目标检测模型，支持nano/tiny/s/m/l/x版本，x版本COCO val2017数据集精度51.8%。
 
 - 2021.11.03: PaddleDetection发布[release/2.3版本](https://github.com/PaddlePaddle/PaddleDetection/tree/release/2.3)
@@ -367,7 +367,7 @@
 
 **点击“ ✅ ”即可下载对应模型**
 
-详细信息参考[文档](deploy/pphuman)
+详细信息参考[文档](deploy/pipeline)
 
 </details>
 
diff --git a/README_en.md b/README_en.md
index 7ec9de272d68690b67d3010fdf3c1ca5938e8308..820ed56b97a55a7b749ecd0bebf24b5cf7d4f9e5 100644
--- a/README_en.md
+++ b/README_en.md
@@ -22,7 +22,7 @@ English | [简体中文](README_cn.md)
 
   - Release GPU SOTA object detection series models (s/m/l/x) [PP-YOLOE](configs/ppyoloe), supporting s/m/l/x version, achieving mAP as 51.6% on COCO test dataset and 78.1 FPS on Nvidia V100 by PP-YOLOE-l, supporting AMP training and its training speed is 33% faster than PP-YOLOv2.
   - Release enhanced models of [PP-PicoDet](configs/picodet), including PP-PicoDet-XS model with 0.7M parameters, its mAP promoted ~2% on COCO, inference speed accelerated 63% on CPU, and post-processing integrated into the network to optimize deployment pipeline.
-  - Release real-time human analysis tool [PP-Human](deploy/pphuman), which is based on data from real-life situations, supporting pedestrian detection, attribute recognition, human tracking, multi-camera tracking, human statistics and action recognition.
+  - Release real-time human analysis tool [PP-Human](deploy/pipeline), which is based on data from real-life situations, supporting pedestrian detection, attribute recognition, human tracking, multi-camera tracking, human statistics and action recognition.
   - Release [YOLOX](configs/yolox), supporting nano/tiny/s/m/l/x version, achieving mAP as 51.8% on COCO val dataset by YOLOX-x.
 
 - 2021.11.03: Release [release/2.3](https://github.com/PaddlePaddle/Paddleetection/tree/release/2.3) version. Release mobile object detection model ⚡[PP-PicoDet](configs/picodet), mobile keypoint detection model ⚡[PP-TinyPose](configs/keypoint/tiny_pose)，Real-time tracking system [PP-Tracking](deploy/pptracking). Release object detection models, including [Swin-Transformer](configs/faster_rcnn), [TOOD](configs/tood), [GFL](configs/gfl), release [Sniper](configs/sniper) tiny object detection models and optimized [PP-YOLO-EB](configs/ppyolo) model for EdgeBoard. Release mobile keypoint detection model [Lite HRNet](configs/keypoint).
@@ -335,7 +335,7 @@ The relationship between COCO mAP and FPS on Qualcomm Snapdragon 865 of represen
   - [Pedestrian detection](configs/pedestrian/README.md)
   - [Vehicle detection](configs/vehicle/README.md)
 - Scienario Solution
-  - [Real-Time Human Analysis Tool PP-Human](deploy/pphuman)
+  - [Real-Time Human Analysis Tool PP-Human](deploy/pipeline)
 - Competition Solution
   - [Objects365 2019 Challenge champion model](static/docs/featured_model/champion_model/CACascadeRCNN_en.md)
   - [Best single model of Open Images 2019-Object Detection](static/docs/featured_model/champion_model/OIDV5_BASELINE_MODEL_en.md)
diff --git a/configs/keypoint/README.md b/configs/keypoint/README.md
index 03fa3b9790ad1655bee4257bb23e89ca26908387..9e4bac733f3ecbc65654b1727cbe89ef541f6d7d 100644
--- a/configs/keypoint/README.md
+++ b/configs/keypoint/README.md
@@ -82,7 +82,7 @@ MPII数据集
 场景模型
 | 模型 | 方案 | 输入尺寸 | 精度 | 预测速度 |模型权重 | 部署模型 | 说明|
 | :---- | ---|----- | :--------: | :--------: | :------------: |:------------: |:-------------------: |
-| HRNet-w32 + DarkPose | Top-Down|256x192  |  AP: 87.1 (业务数据集)| 单人2.9ms |[下载链接](https://bj.bcebos.com/v1/paddledet/models/pipeline/dark_hrnet_w32_256x192.pdparams) |[下载链接](https://bj.bcebos.com/v1/paddledet/models/pipeline/dark_hrnet_w32_256x192.zip) | 针对摔倒场景特别优化，该模型应用于[PP-Human](../../deploy/pphuman/README.md) |
+| HRNet-w32 + DarkPose | Top-Down|256x192  |  AP: 87.1 (业务数据集)| 单人2.9ms |[下载链接](https://bj.bcebos.com/v1/paddledet/models/pipeline/dark_hrnet_w32_256x192.pdparams) |[下载链接](https://bj.bcebos.com/v1/paddledet/models/pipeline/dark_hrnet_w32_256x192.zip) | 针对摔倒场景特别优化，该模型应用于[PP-Human](../../deploy/pipeline/README.md) |
 
 
 我们同时推出了基于LiteHRNet（Top-Down）针对移动端设备优化的实时关键点检测模型[PP-TinyPose](./tiny_pose/README.md), 欢迎体验。
diff --git a/configs/keypoint/README_en.md b/configs/keypoint/README_en.md
index 205a7d71844750336ed63a6913dbd5d2ec949973..915c8160ab2df6004bb2bed7c1460c66091ecd1f 100644
--- a/configs/keypoint/README_en.md
+++ b/configs/keypoint/README_en.md
@@ -93,7 +93,7 @@ MPII Dataset
 Model for Scenes
 | Model | Strategy | Input Size | Precision | Inference Speed |Model Weights | Model Inference and Deployment | description|
 | :---- | ---|----- | :--------: | :-------: |:------------: |:------------: |:-------------------: |
-| HRNet-w32 + DarkPose | Top-Down|256x192  |  AP: 87.1 (on internal dataset)| 2.9ms per person |[Link](https://bj.bcebos.com/v1/paddledet/models/pipeline/dark_hrnet_w32_256x192.pdparams) |[Link](https://bj.bcebos.com/v1/paddledet/models/pipeline/dark_hrnet_w32_256x192.zip) | Especially optimized for fall scenarios, the model is applied to [PP-Human](../../deploy/pphuman/README_en.md) |
+| HRNet-w32 + DarkPose | Top-Down|256x192  |  AP: 87.1 (on internal dataset)| 2.9ms per person |[Link](https://bj.bcebos.com/v1/paddledet/models/pipeline/dark_hrnet_w32_256x192.pdparams) |[Link](https://bj.bcebos.com/v1/paddledet/models/pipeline/dark_hrnet_w32_256x192.zip) | Especially optimized for fall scenarios, the model is applied to [PP-Human](../../deploy/pipeline/README_en.md) |
 
 
 We also release [PP-TinyPose](./tiny_pose/README_en.md), a real-time keypoint detection model optimized for mobile devices. Welcome to experience.
diff --git a/deploy/pphuman/README.md b/deploy/pipeline/README.md
similarity index 100%
rename from deploy/pphuman/README.md
rename to deploy/pipeline/README.md
diff --git a/deploy/pphuman/__init__.py b/deploy/pipeline/__init__.py
similarity index 100%
rename from deploy/pphuman/__init__.py
rename to deploy/pipeline/__init__.py
diff --git a/deploy/pphuman/config/infer_cfg_pphuman.yml b/deploy/pipeline/config/infer_cfg_pphuman.yml
similarity index 95%
rename from deploy/pphuman/config/infer_cfg_pphuman.yml
rename to deploy/pipeline/config/infer_cfg_pphuman.yml
index ed5fb453dc3a19bd19bf7ed0803d99cecc1f7e39..60d986d8534373c0622b11b187c839962d0c1eef 100644
--- a/deploy/pphuman/config/infer_cfg_pphuman.yml
+++ b/deploy/pipeline/config/infer_cfg_pphuman.yml
@@ -10,7 +10,7 @@ DET:
 
 MOT:
   model_dir: output_inference/mot_ppyoloe_l_36e_pipeline/
-  tracker_config: deploy/pphuman/config/tracker_config.yml
+  tracker_config: deploy/pipeline/config/tracker_config.yml
   batch_size: 1
   basemode: "idbased"
   enable: False
diff --git a/deploy/pphuman/config/infer_cfg_ppvehicle.yml b/deploy/pipeline/config/infer_cfg_ppvehicle.yml
similarity index 86%
rename from deploy/pphuman/config/infer_cfg_ppvehicle.yml
rename to deploy/pipeline/config/infer_cfg_ppvehicle.yml
index 0177a8d22426693bcb80df3f5d0bdc88ddb67600..170266510518aca845ac3012e56c6be81b21772d 100644
--- a/deploy/pphuman/config/infer_cfg_ppvehicle.yml
+++ b/deploy/pipeline/config/infer_cfg_ppvehicle.yml
@@ -8,7 +8,7 @@ DET:
 
 MOT:
   model_dir: output_inference/mot_ppyoloe_l_36e_ppvehicle/
-  tracker_config: deploy/pphuman/config/tracker_config.yml
+  tracker_config: deploy/pipeline/config/tracker_config.yml
   batch_size: 1
   basemode: "idbased"
   enable: False
@@ -20,7 +20,7 @@ VEHICLE_PLATE:
   rec_model_dir: output_inference/ch_PP-OCRv3_rec_infer/
   rec_image_shape: [3, 48, 320]
   rec_batch_num: 6
-  word_dict_path: deploy/pphuman/ppvehicle/rec_word_dict.txt
+  word_dict_path: deploy/pipeline/ppvehicle/rec_word_dict.txt
   basemode: "idbased"
   enable: False
 
diff --git a/deploy/pphuman/config/tracker_config.yml b/deploy/pipeline/config/tracker_config.yml
similarity index 100%
rename from deploy/pphuman/config/tracker_config.yml
rename to deploy/pipeline/config/tracker_config.yml
diff --git a/deploy/pphuman/datacollector.py b/deploy/pipeline/datacollector.py
similarity index 97%
rename from deploy/pphuman/datacollector.py
rename to deploy/pipeline/datacollector.py
index 50037b95f514468f788e340a24892a9d38d9d573..fa12fc318022b686c6c73595983b880e29625b66 100644
--- a/deploy/pphuman/datacollector.py
+++ b/deploy/pipeline/datacollector.py
@@ -46,7 +46,7 @@ class Result(object):
 
 class DataCollector(object):
     """
-  DataCollector of pphuman Pipeline, collect results in every frames and assign it to each track ids.
+  DataCollector of Pipeline, collect results in every frames and assign it to each track ids.
   mainly used in mtmct.
   
   data struct:
diff --git a/deploy/pphuman/docs/images/action.gif b/deploy/pipeline/docs/images/action.gif
similarity index 100%
rename from deploy/pphuman/docs/images/action.gif
rename to deploy/pipeline/docs/images/action.gif
diff --git a/deploy/pphuman/docs/images/attribute.gif b/deploy/pipeline/docs/images/attribute.gif
similarity index 100%
rename from deploy/pphuman/docs/images/attribute.gif
rename to deploy/pipeline/docs/images/attribute.gif
diff --git a/deploy/pphuman/docs/images/c1.gif b/deploy/pipeline/docs/images/c1.gif
similarity index 100%
rename from deploy/pphuman/docs/images/c1.gif
rename to deploy/pipeline/docs/images/c1.gif
diff --git a/deploy/pphuman/docs/images/c2.gif b/deploy/pipeline/docs/images/c2.gif
similarity index 100%
rename from deploy/pphuman/docs/images/c2.gif
rename to deploy/pipeline/docs/images/c2.gif
diff --git a/deploy/pphuman/docs/images/fight_demo.gif b/deploy/pipeline/docs/images/fight_demo.gif
similarity index 100%
rename from deploy/pphuman/docs/images/fight_demo.gif
rename to deploy/pipeline/docs/images/fight_demo.gif
diff --git a/deploy/pphuman/docs/images/mot.gif b/deploy/pipeline/docs/images/mot.gif
similarity index 100%
rename from deploy/pphuman/docs/images/mot.gif
rename to deploy/pipeline/docs/images/mot.gif
diff --git a/deploy/pphuman/docs/tutorials/QUICK_STARTED.md b/deploy/pipeline/docs/tutorials/QUICK_STARTED.md
similarity index 82%
rename from deploy/pphuman/docs/tutorials/QUICK_STARTED.md
rename to deploy/pipeline/docs/tutorials/QUICK_STARTED.md
index 32818c3f2b3d942f74335583167bb2e3e4848164..2aceb01d8c541133e9e48b43d316a1f8cfc71bf4 100644
--- a/deploy/pphuman/docs/tutorials/QUICK_STARTED.md
+++ b/deploy/pipeline/docs/tutorials/QUICK_STARTED.md
@@ -50,7 +50,7 @@ PP-Human提供了目标检测、属性识别、行为识别、ReID预训练模
 
 ## 三、配置文件说明
 
-PP-Human相关配置位于```deploy/pphuman/config/infer_cfg_pphuman.yml```中，存放模型路径，完成不同功能需要设置不同的任务类型
+PP-Human相关配置位于```deploy/pipeline/config/infer_cfg_pphuman.yml```中，存放模型路径，完成不同功能需要设置不同的任务类型
 
 功能及任务类型对应表单如下：
 
@@ -69,7 +69,7 @@ visual: True
 
 MOT:
   model_dir: output_inference/mot_ppyoloe_l_36e_pipeline/
-  tracker_config: deploy/pphuman/config/tracker_config.yml
+  tracker_config: deploy/pipeline/config/tracker_config.yml
   batch_size: 1
   basemode: "idbased"
   enable: False
@@ -91,23 +91,23 @@ ATTR:
 
 ```
 # 行人检测，指定配置文件路径和测试图片
-python deploy/pphuman/pipeline.py --config deploy/pphuman/config/infer_cfg_pphuman.yml --image_file=test_image.jpg --device=gpu [--run_mode trt_fp16]
+python deploy/pipeline/pipeline.py --config deploy/pipeline/config/infer_cfg_pphuman.yml --image_file=test_image.jpg --device=gpu [--run_mode trt_fp16]
 
-# 行人跟踪，指定配置文件路径和测试视频，在配置文件中```deploy/pphuman/config/infer_cfg_pphuman.yml```中的MOT部分enable设置为```True```
-python deploy/pphuman/pipeline.py --config deploy/pphuman/config/infer_cfg_pphuman.yml --video_file=test_video.mp4 --device=gpu [--run_mode trt_fp16]
+# 行人跟踪，指定配置文件路径和测试视频，在配置文件中```deploy/pipeline/config/infer_cfg_pphuman.yml```中的MOT部分enable设置为```True```
+python deploy/pipeline/pipeline.py --config deploy/pipeline/config/infer_cfg_pphuman.yml --video_file=test_video.mp4 --device=gpu [--run_mode trt_fp16]
 
-# 行人跟踪，指定配置文件路径，模型路径和测试视频，在配置文件中```deploy/pphuman/config/infer_cfg_pphuman.yml```中的MOT部分enable设置为```True```
+# 行人跟踪，指定配置文件路径，模型路径和测试视频，在配置文件中```deploy/pipeline/config/infer_cfg_pphuman.yml```中的MOT部分enable设置为```True```
 # 命令行中指定的模型路径优先级高于配置文件
-python deploy/pphuman/pipeline.py --config deploy/pphuman/config/infer_cfg_pphuman.yml --video_file=test_video.mp4 --device=gpu --model_dir det=ppyoloe/ [--run_mode trt_fp16]
+python deploy/pipeline/pipeline.py --config deploy/pipeline/config/infer_cfg_pphuman.yml --video_file=test_video.mp4 --device=gpu --model_dir det=ppyoloe/ [--run_mode trt_fp16]
 
-# 行人属性识别，指定配置文件路径和测试视频，在配置文件中```deploy/pphuman/config/infer_cfg_pphuman.yml```中的ATTR部分enable设置为```True```
-python deploy/pphuman/pipeline.py --config deploy/pphuman/config/infer_cfg_pphuman.yml --video_file=test_video.mp4 --device=gpu [--run_mode trt_fp16]
+# 行人属性识别，指定配置文件路径和测试视频，在配置文件中```deploy/pipeline/config/infer_cfg_pphuman.yml```中的ATTR部分enable设置为```True```
+python deploy/pipeline/pipeline.py --config deploy/pipeline/config/infer_cfg_pphuman.yml --video_file=test_video.mp4 --device=gpu [--run_mode trt_fp16]
 
-# 行为识别，指定配置文件路径和测试视频，在配置文件中```deploy/pphuman/config/infer_cfg_pphuman.yml```中的SKELETON_ACTION部分enable设置为```True```
-python deploy/pphuman/pipeline.py --config deploy/pphuman/config/infer_cfg_pphuman.yml --video_file=test_video.mp4 --device=gpu [--run_mode trt_fp16]
+# 行为识别，指定配置文件路径和测试视频，在配置文件中```deploy/pipeline/config/infer_cfg_pphuman.yml```中的SKELETON_ACTION部分enable设置为```True```
+python deploy/pipeline/pipeline.py --config deploy/pipeline/config/infer_cfg_pphuman.yml --video_file=test_video.mp4 --device=gpu [--run_mode trt_fp16]
 
-# 行人跨境跟踪，指定配置文件路径和测试视频列表文件夹，在配置文件中```deploy/pphuman/config/infer_cfg_pphuman.yml```中的REID部分enable设置为```True```
-python deploy/pphuman/pipeline.py --config deploy/pphuman/config/infer_cfg_pphuman.yml --video_dir=mtmct_dir/ --device=gpu [--run_mode trt_fp16]
+# 行人跨境跟踪，指定配置文件路径和测试视频列表文件夹，在配置文件中```deploy/pipeline/config/infer_cfg_pphuman.yml```中的REID部分enable设置为```True```
+python deploy/pipeline/pipeline.py --config deploy/pipeline/config/infer_cfg_pphuman.yml --video_dir=mtmct_dir/ --device=gpu [--run_mode trt_fp16]
 ```
 
 ### 4.1 参数说明
diff --git a/deploy/pphuman/docs/tutorials/action.md b/deploy/pipeline/docs/tutorials/action.md
similarity index 93%
rename from deploy/pphuman/docs/tutorials/action.md
rename to deploy/pipeline/docs/tutorials/action.md
index 8c403784c6a9789b21e5d4a7383e78382f5d6b4b..8b096a85e6b19e2c532915321624007c39ff2977 100644
--- a/deploy/pphuman/docs/tutorials/action.md
+++ b/deploy/pipeline/docs/tutorials/action.md
@@ -45,16 +45,16 @@ SKELETON_ACTION:
 1. 从上表链接中下载模型并解压到```./output_inference```路径下。
 2. 目前行为识别模块仅支持视频输入，设置infer_cfg_pphuman.yml中`SKELETON_ACTION`的enable: True, 然后启动命令如下：
 ```python
-python deploy/pphuman/pipeline.py --config deploy/pphuman/config/infer_cfg_pphuman.yml \
+python deploy/pipeline/pipeline.py --config deploy/pipeline/config/infer_cfg_pphuman.yml \
                                                    --video_file=test_video.mp4 \
                                                    --device=gpu \
 ```
 3. 若修改模型路径，有以下两种方式：
 
-    - ```./deploy/pphuman/config/infer_cfg_pphuman.yml```下可以配置不同模型路径，关键点模型和摔倒行为识别模型分别对应`KPT`和`SKELETON_ACTION`字段，修改对应字段下的路径为实际期望的路径即可。
+    - ```./deploy/pipeline/config/infer_cfg_pphuman.yml```下可以配置不同模型路径，关键点模型和摔倒行为识别模型分别对应`KPT`和`SKELETON_ACTION`字段，修改对应字段下的路径为实际期望的路径即可。
     - 命令行中增加`--model_dir`修改模型路径：
 ```python
-python deploy/pphuman/pipeline.py --config deploy/pphuman/config/infer_cfg_pphuman.yml \
+python deploy/pipeline/pipeline.py --config deploy/pipeline/config/infer_cfg_pphuman.yml \
                                                    --video_file=test_video.mp4 \
                                                    --device=gpu \
                                                    --model_dir kpt=./dark_hrnet_w32_256x192 action=./STGCN
@@ -93,10 +93,10 @@ python deploy/pphuman/pipeline.py --config deploy/pphuman/config/infer_cfg_pphum
 ### 使用方法
 1. 从上表链接中下载预测部署模型并解压到`./output_inference`路径下；
 2. 修改解压后`ppTSM`文件夹中的文件名称为`model.pdiparams、model.pdiparams.info和model.pdmodel`；
-3. 修改配置文件`deploy/pphuman/config/infer_cfg_pphuman.yml`中`VIDEO_ACTION`下的`enable`为`True`；
+3. 修改配置文件`deploy/pipeline/config/infer_cfg_pphuman.yml`中`VIDEO_ACTION`下的`enable`为`True`；
 4. 输入视频，启动命令如下：
 ```
-python deploy/pphuman/pipeline.py --config deploy/pphuman/config/infer_cfg_pphuman.yml \
+python deploy/pipeline/pipeline.py --config deploy/pipeline/config/infer_cfg_pphuman.yml \
                                                    --video_file=test_video.mp4 \
                                                    --device=gpu
 ```
diff --git a/deploy/pphuman/docs/tutorials/action_en.md b/deploy/pipeline/docs/tutorials/action_en.md
similarity index 93%
rename from deploy/pphuman/docs/tutorials/action_en.md
rename to deploy/pipeline/docs/tutorials/action_en.md
index 0d059a3675c179ae954092250e55482c36d02a34..e534ec7cdbce562753d866ea354d66d765c4469a 100644
--- a/deploy/pphuman/docs/tutorials/action_en.md
+++ b/deploy/pipeline/docs/tutorials/action_en.md
@@ -55,19 +55,19 @@ SKELETON_ACTION:
 - Now the only available input is the video input in the action recognition module. set the "enable: True" in SKELETON_ACTION of infer_cfg_pphuman.yml. And then run the command:
 
   ```python
-  python deploy/pphuman/pipeline.py --config deploy/pphuman/config/infer_cfg_pphuman.yml \
+  python deploy/pipeline/pipeline.py --config deploy/pipeline/config/infer_cfg_pphuman.yml \
                                                      --video_file=test_video.mp4 \
                                                      --device=gpu
   ```
 
 - There are two ways to modify the model path:
 
-  - In ```./deploy/pphuman/config/infer_cfg_pphuman.yml```, you can configurate different model paths，which is proper only if you match keypoint models and action recognition models with the fields of `KPT` and `SKELETON_ACTION` respectively, and modify the corresponding path of each field into the expected path.
+  - In ```./deploy/pipeline/config/infer_cfg_pphuman.yml```, you can configurate different model paths，which is proper only if you match keypoint models and action recognition models with the fields of `KPT` and `SKELETON_ACTION` respectively, and modify the corresponding path of each field into the expected path.
 
   - Add `--model_dir` in the command line to revise the model path：
 
     ```python
-    python deploy/pphuman/pipeline.py --config deploy/pphuman/config/infer_cfg_pphuman.yml \
+    python deploy/pipeline/pipeline.py --config deploy/pipeline/config/infer_cfg_pphuman.yml \
                                                        --video_file=test_video.mp4 \
                                                        --device=gpu \
                                                        --model_dir kpt=./dark_hrnet_w32_256x192 action=./STGCN
diff --git a/deploy/pphuman/docs/tutorials/attribute.md b/deploy/pipeline/docs/tutorials/attribute.md
similarity index 90%
rename from deploy/pphuman/docs/tutorials/attribute.md
rename to deploy/pipeline/docs/tutorials/attribute.md
index 8fe4ac3497b45259a24420d3a67a1030a464e3eb..067a017e7dbdf65ae21eba83dbe4918daef94a0a 100644
--- a/deploy/pphuman/docs/tutorials/attribute.md
+++ b/deploy/pipeline/docs/tutorials/attribute.md
@@ -18,22 +18,22 @@
 1. 从上表链接中下载模型并解压到```./output_inference```路径下，并且设置infer_cfg_pphuman.yml中`ATTR`的enable: True
 2. 图片输入时，启动命令如下
 ```python
-python deploy/pphuman/pipeline.py --config deploy/pphuman/config/infer_cfg_pphuman.yml \
+python deploy/pipeline/pipeline.py --config deploy/pipeline/config/infer_cfg_pphuman.yml \
                                                    --image_file=test_image.jpg \
                                                    --device=gpu \
 ```
 3. 视频输入时，启动命令如下
 ```python
-python deploy/pphuman/pipeline.py --config deploy/pphuman/config/infer_cfg_pphuman.yml \
+python deploy/pipeline/pipeline.py --config deploy/pipeline/config/infer_cfg_pphuman.yml \
                                                    --video_file=test_video.mp4 \
                                                    --device=gpu \
 ```
 4. 若修改模型路径，有以下两种方式：
 
-    - ```./deploy/pphuman/config/infer_cfg_pphuman.yml```下可以配置不同模型路径，属性识别模型修改ATTR字段下配置
+    - ```./deploy/pipeline/config/infer_cfg_pphuman.yml```下可以配置不同模型路径，属性识别模型修改ATTR字段下配置
     - **(推荐)**命令行中增加`--model_dir`修改模型路径：
 ```python
-python deploy/pphuman/pipeline.py --config deploy/pphuman/config/infer_cfg_pphuman.yml \
+python deploy/pipeline/pipeline.py --config deploy/pipeline/config/infer_cfg_pphuman.yml \
                                                    --video_file=test_video.mp4 \
                                                    --device=gpu \
                                                    --model_dir det=ppyoloe/
diff --git a/deploy/pphuman/docs/tutorials/attribute_en.md b/deploy/pipeline/docs/tutorials/attribute_en.md
similarity index 90%
rename from deploy/pphuman/docs/tutorials/attribute_en.md
rename to deploy/pipeline/docs/tutorials/attribute_en.md
index 00b0e32d41799f174e1f589114aaf9bc6af2cecd..3d8fab989482ffed72f0d510c7ce21bc485506ba 100644
--- a/deploy/pphuman/docs/tutorials/attribute_en.md
+++ b/deploy/pipeline/docs/tutorials/attribute_en.md
@@ -18,22 +18,22 @@ Pedestrian attribute recognition has been widely used in the intelligent communi
 1. Download the model from the link in the above table, and unzip it to```./output_inference```, and set the "enable: True" in ATTR of infer_cfg_pphuman.yml
 2. When inputting the image, run the command as follows:
 ```python
-python deploy/pphuman/pipeline.py --config deploy/pphuman/config/infer_cfg_pphuman.yml \
+python deploy/pipeline/pipeline.py --config deploy/pipeline/config/infer_cfg_pphuman.yml \
                                                    --image_file=test_image.jpg \
                                                    --device=gpu \
 ```
 3. When inputting the video, run the command as follows:
 ```python
-python deploy/pphuman/pipeline.py --config deploy/pphuman/config/infer_cfg_pphuman.yml \
+python deploy/pipeline/pipeline.py --config deploy/pipeline/config/infer_cfg_pphuman.yml \
                                                    --video_file=test_video.mp4 \
                                                    --device=gpu \
 ```
 4. If you want to change the model path, there are two methods：
 
-    - In ```./deploy/pphuman/config/infer_cfg_pphuman.yml``` you can configurate different model paths. In attribute recognition models, you can modify the configuration in the field of ATTR.
+    - In ```./deploy/pipeline/config/infer_cfg_pphuman.yml``` you can configurate different model paths. In attribute recognition models, you can modify the configuration in the field of ATTR.
     - Add `--model_dir` in the command line to change the model path：
 ```python
-python deploy/pphuman/pipeline.py --config deploy/pphuman/config/infer_cfg_pphuman.yml \
+python deploy/pipeline/pipeline.py --config deploy/pipeline/config/infer_cfg_pphuman.yml \
                                                    --video_file=test_video.mp4 \
                                                    --device=gpu \
                                                    --model_dir det=ppyoloe/
diff --git a/deploy/pphuman/docs/tutorials/mot.md b/deploy/pipeline/docs/tutorials/mot.md
similarity index 88%
rename from deploy/pphuman/docs/tutorials/mot.md
rename to deploy/pipeline/docs/tutorials/mot.md
index 04d05616ed4481025e3d7e6d45442f23d420162b..3e64fb6e1370564c5ca8b9a81c16e00c688dc761 100644
--- a/deploy/pphuman/docs/tutorials/mot.md
+++ b/deploy/pipeline/docs/tutorials/mot.md
@@ -17,22 +17,22 @@
 1. 从上表链接中下载模型并解压到```./output_inference```路径下
 2. 图片输入时，是纯检测任务，启动命令如下
 ```python
-python deploy/pphuman/pipeline.py --config deploy/pphuman/config/infer_cfg_pphuman.yml \
+python deploy/pipeline/pipeline.py --config deploy/pipeline/config/infer_cfg_pphuman.yml \
                                                    --image_file=test_image.jpg \
                                                    --device=gpu
 ```
 3. 视频输入时，是跟踪任务，注意首先设置infer_cfg_pphuman.yml中的MOT配置的enable=True，然后启动命令如下
 ```python
-python deploy/pphuman/pipeline.py --config deploy/pphuman/config/infer_cfg_pphuman.yml \
+python deploy/pipeline/pipeline.py --config deploy/pipeline/config/infer_cfg_pphuman.yml \
                                                    --video_file=test_video.mp4 \
                                                    --device=gpu
 ```
 4. 若修改模型路径，有以下两种方式：
 
-    - ```./deploy/pphuman/config/infer_cfg_pphuman.yml```下可以配置不同模型路径，检测和跟踪模型分别对应`DET`和`MOT`字段，修改对应字段下的路径为实际期望的路径即可。
+    - ```./deploy/pipeline/config/infer_cfg_pphuman.yml```下可以配置不同模型路径，检测和跟踪模型分别对应`DET`和`MOT`字段，修改对应字段下的路径为实际期望的路径即可。
     - 命令行中增加`--model_dir`修改模型路径：
 ```python
-python deploy/pphuman/pipeline.py --config deploy/pphuman/config/infer_cfg_pphuman.yml \
+python deploy/pipeline/pipeline.py --config deploy/pipeline/config/infer_cfg_pphuman.yml \
                                                    --video_file=test_video.mp4 \
                                                    --device=gpu \
                                                    --do_entrance_counting \
diff --git a/deploy/pphuman/docs/tutorials/mot_en.md b/deploy/pipeline/docs/tutorials/mot_en.md
similarity index 87%
rename from deploy/pphuman/docs/tutorials/mot_en.md
rename to deploy/pipeline/docs/tutorials/mot_en.md
index 86a64991ecc88585a7aba6fc070e968cef457bf3..05ad1f7fd84d2a0fed2b2aaa2464ae39bd682035 100644
--- a/deploy/pphuman/docs/tutorials/mot_en.md
+++ b/deploy/pipeline/docs/tutorials/mot_en.md
@@ -17,23 +17,23 @@ Pedestrian detection and tracking is widely used in the intelligent community, i
 1. Download models from the links of the above table and unizp them to ```./output_inference```.
 2. When use the image as input, it's a detection task, the start command is as follows:
 ```python
-python deploy/pphuman/pipeline.py --config deploy/pphuman/config/infer_cfg_pphuman.yml \
+python deploy/pipeline/pipeline.py --config deploy/pipeline/config/infer_cfg_pphuman.yml \
                                                    --image_file=test_image.jpg \
                                                    --device=gpu
 ```
 3. When use the video as input, it's a tracking task, first you should set the "enable: True" in MOT of infer_cfg_pphuman.yml, and then the start command is as follows:
 ```python
-python deploy/pphuman/pipeline.py --config deploy/pphuman/config/infer_cfg_pphuman.yml \
+python deploy/pipeline/pipeline.py --config deploy/pipeline/config/infer_cfg_pphuman.yml \
                                                    --video_file=test_video.mp4 \
                                                    --device=gpu
 ```
 4. There are two ways to modify the model path:
 
-     - In `./deploy/pphuman/config/infer_cfg_pphuman.yml`, you can configurate different model paths，which is proper only if you match keypoint models and action recognition models with the fields of `DET` and `MOT` respectively, and modify the corresponding path of each field into the expected path.
+     - In `./deploy/pipeline/config/infer_cfg_pphuman.yml`, you can configurate different model paths，which is proper only if you match keypoint models and action recognition models with the fields of `DET` and `MOT` respectively, and modify the corresponding path of each field into the expected path.
     - Add `--model_dir` in the command line to revise the model path:
 
 ```python
-python deploy/pphuman/pipeline.py --config deploy/pphuman/config/infer_cfg_pphuman.yml \
+python deploy/pipeline/pipeline.py --config deploy/pipeline/config/infer_cfg_pphuman.yml \
                                                    --video_file=test_video.mp4 \
                                                    --device=gpu \
                                                    --model_dir det=ppyoloe/
diff --git a/deploy/pphuman/docs/tutorials/mtmct.md b/deploy/pipeline/docs/tutorials/mtmct.md
similarity index 90%
rename from deploy/pphuman/docs/tutorials/mtmct.md
rename to deploy/pipeline/docs/tutorials/mtmct.md
index 16a9fda8e0cb2d03bd548674ec5402e32fd0416a..54e407ab873771d0b393496a0c02b700954e90df 100644
--- a/deploy/pphuman/docs/tutorials/mtmct.md
+++ b/deploy/pipeline/docs/tutorials/mtmct.md
@@ -11,14 +11,14 @@ PP-Human跨镜头跟踪模块主要目的在于提供一套简洁、高效的跨
 
 2. 跨镜头跟踪模式下，要求输入的多个视频放在同一目录下，同时开启infer_cfg_pphuman.yml 中的REID选择中的enable=True, 命令如下：
 ```python
-python3 deploy/pphuman/pipeline.py --config deploy/pphuman/config/infer_cfg_pphuman.yml --video_dir=[your_video_file_directory] --device=gpu
+python3 deploy/pipeline/pipeline.py --config deploy/pipeline/config/infer_cfg_pphuman.yml --video_dir=[your_video_file_directory] --device=gpu
 ```
 
-3. 相关配置在`./deploy/pphuman/config/infer_cfg_pphuman.yml`文件中修改：
+3. 相关配置在`./deploy/pipeline/config/infer_cfg_pphuman.yml`文件中修改：
 
 ```python
-python3 deploy/pphuman/pipeline.py
-        --config deploy/pphuman/config/infer_cfg_pphuman.yml
+python3 deploy/pipeline/pipeline.py
+        --config deploy/pipeline/config/infer_cfg_pphuman.yml
         --video_dir=[your_video_file_directory]
         --device=gpu
         --model_dir reid=reid_best/
diff --git a/deploy/pphuman/docs/tutorials/mtmct_en.md b/deploy/pipeline/docs/tutorials/mtmct_en.md
similarity index 90%
rename from deploy/pphuman/docs/tutorials/mtmct_en.md
rename to deploy/pipeline/docs/tutorials/mtmct_en.md
index 117481a21a8653ef0cc87641d06c39733455079b..756edec367b60c872bf8881e7731fff5f9c77d1f 100644
--- a/deploy/pphuman/docs/tutorials/mtmct_en.md
+++ b/deploy/pipeline/docs/tutorials/mtmct_en.md
@@ -11,14 +11,14 @@ The MTMCT module of PP-Human aims to provide a multi-target multi-camera pipleli
 
 2. In the MTMCT mode, input videos are required to be put in the same directory. set the REID "enable: True" in the infer_cfg_pphuman.yml. The command line is:
 ```python
-python3 deploy/pphuman/pipeline.py --config deploy/pphuman/config/infer_cfg_pphuman.yml --video_dir=[your_video_file_directory] --device=gpu
+python3 deploy/pipeline/pipeline.py --config deploy/pipeline/config/infer_cfg_pphuman.yml --video_dir=[your_video_file_directory] --device=gpu
 ```
 
-3. Configuration can be modified in `./deploy/pphuman/config/infer_cfg_pphuman.yml`.
+3. Configuration can be modified in `./deploy/pipeline/config/infer_cfg_pphuman.yml`.
 
 ```python
-python3 deploy/pphuman/pipeline.py
-        --config deploy/pphuman/config/infer_cfg_pphuman.yml
+python3 deploy/pipeline/pipeline.py
+        --config deploy/pipeline/config/infer_cfg_pphuman.yml
         --video_dir=[your_video_file_directory]
         --device=gpu
         --model_dir reid=reid_best/
diff --git a/deploy/pphuman/pipe_utils.py b/deploy/pipeline/pipe_utils.py
similarity index 100%
rename from deploy/pphuman/pipe_utils.py
rename to deploy/pipeline/pipe_utils.py
diff --git a/deploy/pphuman/pipeline.py b/deploy/pipeline/pipeline.py
similarity index 99%
rename from deploy/pphuman/pipeline.py
rename to deploy/pipeline/pipeline.py
index c87a7c6ba2ab5cfa5fa721a5107cc41e643b6981..886d1e3355494eb3a377ce3dcb9446380b2ad2de 100644
--- a/deploy/pphuman/pipeline.py
+++ b/deploy/pipeline/pipeline.py
@@ -15,34 +15,25 @@
 import os
 import yaml
 import glob
-from collections import defaultdict
-
 import cv2
 import numpy as np
 import math
 import paddle
 import sys
 import copy
-from collections import Sequence
-from reid import ReID
+from collections import Sequence, defaultdict
 from datacollector import DataCollector, Result
-from mtmct import mtmct_process
 
 # add deploy path of PadleDetection to sys.path
 parent_path = os.path.abspath(os.path.join(__file__, *(['..'] * 2)))
 sys.path.insert(0, parent_path)
 
+from pipe_utils import argsparser, print_arguments, merge_cfg, PipeTimer
+from pipe_utils import get_test_images, crop_image_with_det, crop_image_with_mot, parse_mot_res, parse_mot_keypoint
+
 from python.infer import Detector, DetectorPicoDet
-from python.attr_infer import AttrDetector
 from python.keypoint_infer import KeyPointDetector
 from python.keypoint_postprocess import translate_to_ori_images
-
-from python.video_action_infer import VideoActionRecognizer
-from python.action_infer import SkeletonActionRecognizer, DetActionRecognizer, ClsActionRecognizer
-from python.action_utils import KeyPointBuff, ActionVisualHelper
-
-from pipe_utils import argsparser, print_arguments, merge_cfg, PipeTimer
-from pipe_utils import get_test_images, crop_image_with_det, crop_image_with_mot, parse_mot_res, parse_mot_keypoint
 from python.preprocess import decode_image, ShortSizeScale
 from python.visualize import visualize_box_mask, visualize_attr, visualize_pose, visualize_action, visualize_vehicleplate
 
@@ -50,6 +41,13 @@ from pptracking.python.mot_sde_infer import SDE_Detector
 from pptracking.python.mot.visualize import plot_tracking_dict
 from pptracking.python.mot.utils import flow_statistic
 
+from pphuman.attr_infer import AttrDetector
+from pphuman.video_action_infer import VideoActionRecognizer
+from pphuman.action_infer import SkeletonActionRecognizer, DetActionRecognizer, ClsActionRecognizer
+from pphuman.action_utils import KeyPointBuff, ActionVisualHelper
+from pphuman.reid import ReID
+from pphuman.mtmct import mtmct_process
+
 from ppvehicle.vehicle_plate import PlateRecognizer
 from ppvehicle.vehicle_attr import VehicleAttr
 
diff --git a/deploy/python/action_infer.py b/deploy/pipeline/pphuman/action_infer.py
similarity index 99%
rename from deploy/python/action_infer.py
rename to deploy/pipeline/pphuman/action_infer.py
index 4c61f06c265b48112b3c9e3da3eb105688546851..563d6f7f5db910a23288d7b683df6c88f5dd2150 100644
--- a/deploy/python/action_infer.py
+++ b/deploy/pipeline/pphuman/action_infer.py
@@ -28,9 +28,9 @@ parent_path = os.path.abspath(os.path.join(__file__, *(['..'] * 2)))
 sys.path.insert(0, parent_path)
 
 from paddle.inference import Config, create_predictor
-from utils import argsparser, Timer, get_current_memory_mb
-from benchmark_utils import PaddleInferBenchmark
-from infer import Detector, print_arguments
+from python.utils import argsparser, Timer, get_current_memory_mb
+from python.benchmark_utils import PaddleInferBenchmark
+from python.infer import Detector, print_arguments
 from attr_infer import AttrDetector
 
 
diff --git a/deploy/python/action_utils.py b/deploy/pipeline/pphuman/action_utils.py
similarity index 100%
rename from deploy/python/action_utils.py
rename to deploy/pipeline/pphuman/action_utils.py
diff --git a/deploy/python/attr_infer.py b/deploy/pipeline/pphuman/attr_infer.py
similarity index 97%
rename from deploy/python/attr_infer.py
rename to deploy/pipeline/pphuman/attr_infer.py
index ba034639a959a89df9ed49b7256b316ab541773f..f783d5e5c857a374c8bfb241eb5bebd18bf9cd84 100644
--- a/deploy/python/attr_infer.py
+++ b/deploy/pipeline/pphuman/attr_infer.py
@@ -29,11 +29,11 @@ import sys
 parent_path = os.path.abspath(os.path.join(__file__, *(['..'])))
 sys.path.insert(0, parent_path)
 
-from benchmark_utils import PaddleInferBenchmark
-from preprocess import preprocess, Resize, NormalizeImage, Permute, PadStride, LetterBoxResize, WarpAffine
-from visualize import visualize_attr
-from utils import argsparser, Timer, get_current_memory_mb
-from infer import Detector, get_test_images, print_arguments, load_predictor
+from python.benchmark_utils import PaddleInferBenchmark
+from python.preprocess import preprocess, Resize, NormalizeImage, Permute, PadStride, LetterBoxResize, WarpAffine
+from python.visualize import visualize_attr
+from python.utils import argsparser, Timer, get_current_memory_mb
+from python.infer import Detector, get_test_images, print_arguments, load_predictor
 
 from PIL import Image, ImageDraw, ImageFont
 
diff --git a/deploy/pphuman/mtmct.py b/deploy/pipeline/pphuman/mtmct.py
similarity index 99%
rename from deploy/pphuman/mtmct.py
rename to deploy/pipeline/pphuman/mtmct.py
index 7c4d571155e932dc7bb4cf77b10ab5cb6c0be66c..721008f638b543c9f276c0e86877ecbe32eb2c46 100644
--- a/deploy/pphuman/mtmct.py
+++ b/deploy/pipeline/pphuman/mtmct.py
@@ -12,7 +12,6 @@
 # See the License for the specific language governing permissions and
 # limitations under the License.
 
-import motmetrics as mm
 from pptracking.python.mot.visualize import plot_tracking
 import os
 import re
diff --git a/deploy/pphuman/reid.py b/deploy/pipeline/pphuman/reid.py
similarity index 100%
rename from deploy/pphuman/reid.py
rename to deploy/pipeline/pphuman/reid.py
diff --git a/deploy/python/video_action_infer.py b/deploy/pipeline/pphuman/video_action_infer.py
similarity index 96%
rename from deploy/python/video_action_infer.py
rename to deploy/pipeline/pphuman/video_action_infer.py
index 865f47d41ac708f667edfb5a022a0315d69392bd..3a10c033343063b24722f07958c6daa394fc9e0e 100644
--- a/deploy/python/video_action_infer.py
+++ b/deploy/pipeline/pphuman/video_action_infer.py
@@ -29,9 +29,9 @@ parent_path = os.path.abspath(os.path.join(__file__, *(['..'] * 2)))
 sys.path.insert(0, parent_path)
 
 from paddle.inference import Config, create_predictor
-from utils import argsparser, Timer, get_current_memory_mb
-from benchmark_utils import PaddleInferBenchmark
-from infer import Detector, print_arguments
+from python.utils import argsparser, Timer, get_current_memory_mb
+from python.benchmark_utils import PaddleInferBenchmark
+from python.infer import Detector, print_arguments
 from video_action_preprocess import VideoDecoder, Sampler, Scale, CenterCrop, Normalization, Image2Array
 
 
@@ -96,8 +96,8 @@ class VideoActionRecognizer(object):
 
         self.recognize_times = Timer()
 
-        model_file_path = os.path.join(model_dir, "model.pdmodel")
-        params_file_path = os.path.join(model_dir, "model.pdiparams")
+        model_file_path = os.path.join(model_dir, "ppTSM.pdmodel")
+        params_file_path = os.path.join(model_dir, "ppTSM.pdiparams")
         self.config = Config(model_file_path, params_file_path)
 
         if device == "GPU" or device == "gpu":
diff --git a/deploy/python/video_action_preprocess.py b/deploy/pipeline/pphuman/video_action_preprocess.py
similarity index 100%
rename from deploy/python/video_action_preprocess.py
rename to deploy/pipeline/pphuman/video_action_preprocess.py
diff --git a/deploy/pphuman/ppvehicle/rec_word_dict.txt b/deploy/pipeline/ppvehicle/rec_word_dict.txt
similarity index 100%
rename from deploy/pphuman/ppvehicle/rec_word_dict.txt
rename to deploy/pipeline/ppvehicle/rec_word_dict.txt
diff --git a/deploy/pphuman/ppvehicle/vehicle_attr.py b/deploy/pipeline/ppvehicle/vehicle_attr.py
similarity index 96%
rename from deploy/pphuman/ppvehicle/vehicle_attr.py
rename to deploy/pipeline/ppvehicle/vehicle_attr.py
index e937fade73ed375e3c713937236644eedbfde003..82db40b3189571452b36a870049b5c17bdaf554d 100644
--- a/deploy/pphuman/ppvehicle/vehicle_attr.py
+++ b/deploy/pipeline/ppvehicle/vehicle_attr.py
@@ -28,10 +28,10 @@ parent_path = os.path.abspath(os.path.join(__file__, *(['..'] * 3)))
 sys.path.insert(0, parent_path)
 
 from paddle.inference import Config, create_predictor
-from utils import argsparser, Timer, get_current_memory_mb
-from benchmark_utils import PaddleInferBenchmark
+from python.utils import argsparser, Timer, get_current_memory_mb
+from python.benchmark_utils import PaddleInferBenchmark
 from python.infer import Detector, print_arguments
-from python.attr_infer import AttrDetector
+from pipeline.pphuman.attr_infer import AttrDetector
 
 
 class VehicleAttr(AttrDetector):
diff --git a/deploy/pphuman/ppvehicle/vehicle_plate.py b/deploy/pipeline/ppvehicle/vehicle_plate.py
similarity index 89%
rename from deploy/pphuman/ppvehicle/vehicle_plate.py
rename to deploy/pipeline/ppvehicle/vehicle_plate.py
index 5e0c5af95958fbd9dea3996b58df2762ed452bf8..42e5ba4b0a1cb1bfcd7fb1422cf8b700d904294c 100644
--- a/deploy/pphuman/ppvehicle/vehicle_plate.py
+++ b/deploy/pipeline/ppvehicle/vehicle_plate.py
@@ -29,9 +29,9 @@ sys.path.insert(0, parent_path)
 
 from python.infer import get_test_images
 from python.preprocess import preprocess, NormalizeImage, Permute, Resize_Mult32
-from pphuman.ppvehicle.vehicle_plateutils import create_predictor, get_infer_gpuid, get_rotate_crop_image, draw_boxes
-from pphuman.ppvehicle.vehicleplate_postprocess import build_post_process
-from pphuman.pipe_utils import merge_cfg, print_arguments, argsparser
+from pipeline.ppvehicle.vehicle_plateutils import create_predictor, get_infer_gpuid, get_rotate_crop_image, draw_boxes
+from pipeline.ppvehicle.vehicleplate_postprocess import build_post_process
+from pipeline.pipe_utils import merge_cfg, print_arguments, argsparser
 
 
 class PlateDetector(object):
@@ -62,18 +62,17 @@ class PlateDetector(object):
         self.predictor, self.input_tensor, self.output_tensors, self.config = create_predictor(
             args, cfg, 'det')
 
-    def preprocess(self, image_list):
+    def preprocess(self, im_path):
         preprocess_ops = []
         for op_type, new_op_info in self.pre_process_list.items():
             preprocess_ops.append(eval(op_type)(**new_op_info))
 
         input_im_lst = []
         input_im_info_lst = []
-        for im_path in image_list:
-            im, im_info = preprocess(im_path, preprocess_ops)
-            input_im_lst.append(im)
-            input_im_info_lst.append(im_info['im_shape'] /
-                                     im_info['scale_factor'])
+
+        im, im_info = preprocess(im_path, preprocess_ops)
+        input_im_lst.append(im)
+        input_im_info_lst.append(im_info['im_shape'] / im_info['scale_factor'])
 
         return np.stack(input_im_lst, axis=0), input_im_info_lst
 
@@ -119,27 +118,28 @@ class PlateDetector(object):
     def predict_image(self, img_list):
         st = time.time()
 
-        img, shape_list = self.preprocess(img_list)
-        if img is None:
-            return None, 0
-
-        self.input_tensor.copy_from_cpu(img)
-        self.predictor.run()
-        outputs = []
-        for output_tensor in self.output_tensors:
-            output = output_tensor.copy_to_cpu()
-            outputs.append(output)
-
-        preds = {}
-        preds['maps'] = outputs[0]
-
-        #self.predictor.try_shrink_memory()
-        post_result = self.postprocess_op(preds, shape_list)
-
         dt_batch_boxes = []
-        for idx in range(len(post_result)):
-            org_shape = img_list[idx].shape
-            dt_boxes = post_result[idx]['points']
+        for image in img_list:
+            img, shape_list = self.preprocess(image)
+            if img is None:
+                return None, 0
+
+            self.input_tensor.copy_from_cpu(img)
+            self.predictor.run()
+            outputs = []
+            for output_tensor in self.output_tensors:
+                output = output_tensor.copy_to_cpu()
+                outputs.append(output)
+
+            preds = {}
+            preds['maps'] = outputs[0]
+
+            #self.predictor.try_shrink_memory()
+            post_result = self.postprocess_op(preds, shape_list)
+            # print("post_result length:{}".format(len(post_result)))
+
+            org_shape = image.shape
+            dt_boxes = post_result[0]['points']
             dt_boxes = self.filter_tag_det_res(dt_boxes, org_shape)
             dt_batch_boxes.append(dt_boxes)
 
diff --git a/deploy/pphuman/ppvehicle/vehicle_plateutils.py b/deploy/pipeline/ppvehicle/vehicle_plateutils.py
similarity index 100%
rename from deploy/pphuman/ppvehicle/vehicle_plateutils.py
rename to deploy/pipeline/ppvehicle/vehicle_plateutils.py
diff --git a/deploy/pphuman/ppvehicle/vehicleplate_postprocess.py b/deploy/pipeline/ppvehicle/vehicleplate_postprocess.py
similarity index 100%
rename from deploy/pphuman/ppvehicle/vehicleplate_postprocess.py
rename to deploy/pipeline/ppvehicle/vehicleplate_postprocess.py
diff --git a/deploy/pphuman/tools/clip_video.py b/deploy/pipeline/tools/clip_video.py
similarity index 100%
rename from deploy/pphuman/tools/clip_video.py
rename to deploy/pipeline/tools/clip_video.py
diff --git a/deploy/pphuman/tools/split_fight_train_test_dataset.py b/deploy/pipeline/tools/split_fight_train_test_dataset.py
similarity index 100%
rename from deploy/pphuman/tools/split_fight_train_test_dataset.py
rename to deploy/pipeline/tools/split_fight_train_test_dataset.py
diff --git a/deploy/python/README.md b/deploy/python/README.md
index cb32ad468b58ff1d0cf0f9cb5f625822c8959393..c24aa7be6dd32bb0762d269afd9f59b39d5ab9a8 100644
--- a/deploy/python/README.md
+++ b/deploy/python/README.md
@@ -52,7 +52,7 @@ python deploy/python/det_keypoint_unite_infer.py --det_model_dir=output_inferenc
 ```
 **注意:**
  - 关键点检测模型导出和预测具体可参照[keypoint](../../configs/keypoint/README.md)，可分别在各个模型的文档中查找具体用法；
- - 此目录下的关键点检测部署为基础前向功能，更多关键点检测功能可使用PP-Human项目，参照[pphuman](../pphuman/README.md)；
+ - 此目录下的关键点检测部署为基础前向功能，更多关键点检测功能可使用PP-Human项目，参照[pipeline](../pipeline/README.md)；
 
 
 ### 2.3 多目标跟踪
@@ -70,7 +70,7 @@ python deploy/python/mot_keypoint_unite_infer.py --mot_model_dir=output_inferenc
 
 **注意:**
  - 多目标跟踪模型导出和预测具体可参照[mot]](../../configs/mot/README.md)，可分别在各个模型的文档中查找具体用法；
- - 此目录下的跟踪部署为基础前向功能以及联合关键点部署，更多跟踪功能可使用PP-Human项目，参照[pphuman](../pphuman/README.md)，或PP-Tracking项目(绘制轨迹、出入口流量计数)，参照[pptracking](../pptracking/README.md)；
+ - 此目录下的跟踪部署为基础前向功能以及联合关键点部署，更多跟踪功能可使用PP-Human项目，参照[pipeline](../pipeline/README.md)，或PP-Tracking项目(绘制轨迹、出入口流量计数)，参照[pptracking](../pptracking/README.md)；
 
 
 参数说明如下:
diff --git a/deploy/python/visualize.py b/deploy/python/visualize.py
index 1d50d68c8f3e4a5ac12ef68d5efa73d8bb1d17fb..8f0da233df29ddc617cef99fe415b9ec7ed39efc 100644
--- a/deploy/python/visualize.py
+++ b/deploy/python/visualize.py
@@ -428,7 +428,7 @@ def visualize_vehicleplate(im, results, boxes=None):
             if text == "":
                 continue
             text_w = int(box[2])
-            text_h = int(box[5])
+            text_h = int(box[5] + box[3])
             text_loc = (text_w, text_h)
             cv2.putText(
                 im,
diff --git a/docs/advanced_tutorials/customization/action.md b/docs/advanced_tutorials/customization/action.md
index c7fbe723bf71fc182341f4d65c5b9de9272aa404..a6d70ff10519253502f26fd04d7652b79c3e3c03 100644
--- a/docs/advanced_tutorials/customization/action.md
+++ b/docs/advanced_tutorials/customization/action.md
@@ -55,12 +55,12 @@ python split_fight_train_test_dataset.py "rawframes" 2 0.8
 
 参数说明：“rawframes”为视频帧存放的文件夹；2表示目录结构为两级，第二级表示每个行为对应的子文件夹；0.8表示训练集比例。
 
-其中`split_fight_train_test_dataset.py`文件在`deploy/pphuman/tools`路径下。
+其中`split_fight_train_test_dataset.py`文件在`deploy/pipeline/tools`路径下。
 
 执行完命令后会最终生成fight_train_list.txt和fight_val_list.txt两个文件。打架的标签为1，非打架的标签为0。
 
 #### 视频裁剪
-对于未裁剪的视频，需要先进行裁剪才能用于模型训练，`deploy/pphuman/tools/clip_video.py`中给出了视频裁剪的函数`cut_video`，输入为视频路径，裁剪的起始帧和结束帧以及裁剪后的视频保存路径。
+对于未裁剪的视频，需要先进行裁剪才能用于模型训练，`deploy/pipeline/tools/clip_video.py`中给出了视频裁剪的函数`cut_video`，输入为视频路径，裁剪的起始帧和结束帧以及裁剪后的视频保存路径。
 
 ## 模型优化