generate_proposals_op.cc

/* Copyright (c) 2018 PaddlePaddle Authors. All Rights Reserved.

Licensed under the Apache License, Version 2.0 (the "License");
you may not use this file except in compliance with the License.
You may obtain a copy of the License at

    http://www.apache.org/licenses/LICENSE-2.0

Unless required by applicable law or agreed to in writing, software
distributed under the License is distributed on an "AS IS" BASIS,
WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
See the License for the specific language governing permissions and
limitations under the License. */

#include <cmath>
#include <cstring>
#include <string>
#include <vector>
#include "paddle/fluid/framework/op_registry.h"
#include "paddle/fluid/operators/detail/safe_ref.h"
#include "paddle/fluid/operators/gather.h"
#include "paddle/fluid/operators/math/math_function.h"

namespace paddle {
namespace operators {

using Tensor = framework::Tensor;
using LoDTensor = framework::LoDTensor;

static const double kBBoxClipDefault = std::log(1000.0 / 16.0);

static void AppendProposals(Tensor *dst, int64_t offset, const Tensor &src) {
  auto *out_data = dst->data<void>();
  auto *to_add_data = src.data<void>();
  size_t size_of_t = framework::SizeOfType(src.type());
  offset *= size_of_t;
  std::memcpy(
      reinterpret_cast<void *>(reinterpret_cast<uintptr_t>(out_data) + offset),
      to_add_data, src.numel() * size_of_t);
}

class GenerateProposalsOp : public framework::OperatorWithKernel {
 public:
  using framework::OperatorWithKernel::OperatorWithKernel;

  void InferShape(framework::InferShapeContext *ctx) const override {
    PADDLE_ENFORCE(ctx->HasInput("Scores"), "Input(Scores) shouldn't be null.");
    PADDLE_ENFORCE(ctx->HasInput("BboxDeltas"),
                   "Input(BboxDeltas) shouldn't be null.");
    PADDLE_ENFORCE(ctx->HasInput("ImInfo"), "Input(ImInfo) shouldn't be null.");
    PADDLE_ENFORCE(ctx->HasInput("Anchors"),
                   "Input(Anchors) shouldn't be null.");
    PADDLE_ENFORCE(ctx->HasInput("Variances"),
                   "Input(Variances) shouldn't be null.");

    auto scores_dims = ctx->GetInputDim("Scores");
    auto bbox_deltas_dims = ctx->GetInputDim("BboxDeltas");
    auto im_info_dims = ctx->GetInputDim("ImInfo");
    auto anchors_dims = ctx->GetInputDim("Anchors");
    auto variances_dims = ctx->GetInputDim("Variances");

    ctx->SetOutputDim("RpnRois", {-1, 4});
    ctx->SetOutputDim("RpnRoiProbs", {-1, 1});
  }

 protected:
  framework::OpKernelType GetExpectedKernelType(
      const framework::ExecutionContext &ctx) const override {
    return framework::OpKernelType(
        framework::ToDataType(ctx.Input<Tensor>("Anchors")->type()),
        ctx.device_context());
  }
};

template <class T>
static inline void BoxCoder(const platform::DeviceContext &ctx,
                            Tensor *all_anchors, Tensor *bbox_deltas,
                            Tensor *variances, Tensor *proposals) {
  T *proposals_data = proposals->mutable_data<T>(ctx.GetPlace());

  int64_t row = all_anchors->dims()[0];
  int64_t len = all_anchors->dims()[1];

  auto *bbox_deltas_data = bbox_deltas->data<T>();
  auto *anchor_data = all_anchors->data<T>();
  const T *variances_data = nullptr;
  if (variances) {
    variances_data = variances->data<T>();
  }

  for (int64_t i = 0; i < row; ++i) {
    T anchor_width = anchor_data[i * len + 2] - anchor_data[i * len] + 1.0;
    T anchor_height = anchor_data[i * len + 3] - anchor_data[i * len + 1] + 1.0;

    T anchor_center_x = anchor_data[i * len] + 0.5 * anchor_width;
    T anchor_center_y = anchor_data[i * len + 1] + 0.5 * anchor_height;

    T bbox_center_x = 0, bbox_center_y = 0;
    T bbox_width = 0, bbox_height = 0;

    if (variances) {
      bbox_center_x =
          variances_data[i * len] * bbox_deltas_data[i * len] * anchor_width +
          anchor_center_x;
      bbox_center_y = variances_data[i * len + 1] *
                          bbox_deltas_data[i * len + 1] * anchor_height +
                      anchor_center_y;
      bbox_width = std::exp(std::min<T>(variances_data[i * len + 2] *
                                            bbox_deltas_data[i * len + 2],
                                        kBBoxClipDefault)) *
                   anchor_width;
      bbox_height = std::exp(std::min<T>(variances_data[i * len + 3] *
                                             bbox_deltas_data[i * len + 3],
                                         kBBoxClipDefault)) *
                    anchor_height;
    } else {
      bbox_center_x =
          bbox_deltas_data[i * len] * anchor_width + anchor_center_x;
      bbox_center_y =
          bbox_deltas_data[i * len + 1] * anchor_height + anchor_center_y;
      bbox_width = std::exp(std::min<T>(bbox_deltas_data[i * len + 2],
                                        kBBoxClipDefault)) *
                   anchor_width;
      bbox_height = std::exp(std::min<T>(bbox_deltas_data[i * len + 3],
                                         kBBoxClipDefault)) *
                    anchor_height;
    }

    proposals_data[i * len] = bbox_center_x - bbox_width / 2;
    proposals_data[i * len + 1] = bbox_center_y - bbox_height / 2;
    proposals_data[i * len + 2] = bbox_center_x + bbox_width / 2 - 1;
    proposals_data[i * len + 3] = bbox_center_y + bbox_height / 2 - 1;
  }
  // return proposals;
}

template <class T>
static inline void ClipTiledBoxes(const platform::DeviceContext &ctx,
                                  const Tensor &im_info, Tensor *boxes) {
  T *boxes_data = boxes->mutable_data<T>(ctx.GetPlace());
  const T *im_info_data = im_info.data<T>();
  T zero(0);
  for (int64_t i = 0; i < boxes->numel(); ++i) {
    if (i % 4 == 0) {
      boxes_data[i] =
          std::max(std::min(boxes_data[i], im_info_data[1] - 1), zero);
    } else if (i % 4 == 1) {
      boxes_data[i] =
          std::max(std::min(boxes_data[i], im_info_data[0] - 1), zero);
    } else if (i % 4 == 2) {
      boxes_data[i] =
          std::max(std::min(boxes_data[i], im_info_data[1] - 1), zero);
    } else {
      boxes_data[i] =
          std::max(std::min(boxes_data[i], im_info_data[0] - 1), zero);
    }
  }
}

template <class T>
static inline void FilterBoxes(const platform::DeviceContext &ctx,
                               Tensor *boxes, float min_size,
                               const Tensor &im_info, Tensor *keep) {
  const T *im_info_data = im_info.data<T>();
  T *boxes_data = boxes->mutable_data<T>(ctx.GetPlace());
  T im_scale = im_info_data[2];
  keep->Resize({boxes->dims()[0]});
  min_size = std::max(min_size, 1.0f);
  int *keep_data = keep->mutable_data<int>(ctx.GetPlace());

  int keep_len = 0;
  for (int i = 0; i < boxes->dims()[0]; ++i) {
    T ws = boxes_data[4 * i + 2] - boxes_data[4 * i] + 1;
    T hs = boxes_data[4 * i + 3] - boxes_data[4 * i + 1] + 1;
    T ws_origin_scale =
        (boxes_data[4 * i + 2] - boxes_data[4 * i]) / im_scale + 1;
    T hs_origin_scale =
        (boxes_data[4 * i + 3] - boxes_data[4 * i + 1]) / im_scale + 1;
    T x_ctr = boxes_data[4 * i] + ws / 2;
    T y_ctr = boxes_data[4 * i + 1] + hs / 2;
    if (ws_origin_scale >= min_size && hs_origin_scale >= min_size &&
        x_ctr <= im_info_data[1] && y_ctr <= im_info_data[0]) {
      keep_data[keep_len++] = i;
    }
  }
  keep->Resize({keep_len});
}

template <class T>
static inline std::vector<std::pair<T, int>> GetSortedScoreIndex(
    const std::vector<T> &scores) {
  std::vector<std::pair<T, int>> sorted_indices;
  sorted_indices.reserve(scores.size());
  for (size_t i = 0; i < scores.size(); ++i) {
    sorted_indices.emplace_back(scores[i], i);
  }
  // Sort the score pair according to the scores in descending order
  std::stable_sort(sorted_indices.begin(), sorted_indices.end(),
                   [](const std::pair<T, int> &a, const std::pair<T, int> &b) {
                     return a.first < b.first;
                   });
  return sorted_indices;
}

template <class T>
static inline T BBoxArea(const T *box, bool normalized) {
  if (box[2] < box[0] || box[3] < box[1]) {
    // If coordinate values are is invalid
    // (e.g. xmax < xmin or ymax < ymin), return 0.
    return static_cast<T>(0.);
  } else {
    const T w = box[2] - box[0];
    const T h = box[3] - box[1];
    if (normalized) {
      return w * h;
    } else {
      // If coordinate values are not within range [0, 1].
      return (w + 1) * (h + 1);
    }
  }
}

template <class T>
static inline T JaccardOverlap(const T *box1, const T *box2, bool normalized) {
  if (box2[0] > box1[2] || box2[2] < box1[0] || box2[1] > box1[3] ||
      box2[3] < box1[1]) {
    return static_cast<T>(0.);
  } else {
    const T inter_xmin = std::max(box1[0], box2[0]);
    const T inter_ymin = std::max(box1[1], box2[1]);
    const T inter_xmax = std::min(box1[2], box2[2]);
    const T inter_ymax = std::min(box1[3], box2[3]);
    const T inter_w = std::max(T(0), inter_xmax - inter_xmin + 1);
    const T inter_h = std::max(T(0), inter_ymax - inter_ymin + 1);
    const T inter_area = inter_w * inter_h;
    const T bbox1_area = BBoxArea<T>(box1, normalized);
    const T bbox2_area = BBoxArea<T>(box2, normalized);
    return inter_area / (bbox1_area + bbox2_area - inter_area);
  }
}

template <typename T>
static inline Tensor VectorToTensor(const std::vector<T> &selected_indices,
                                    int selected_num) {
  Tensor keep_nms;
  keep_nms.Resize({selected_num});
  auto *keep_data = keep_nms.mutable_data<T>(platform::CPUPlace());
  for (int i = 0; i < selected_num; ++i) {
    keep_data[i] = selected_indices[i];
  }
  return keep_nms;
}

template <class T>
static inline Tensor NMS(const platform::DeviceContext &ctx, Tensor *bbox,
                         Tensor *scores, T nms_threshold, float eta) {
  PADDLE_ENFORCE_NOT_NULL(bbox);
  int64_t num_boxes = bbox->dims()[0];
  // 4: [xmin ymin xmax ymax]
  int64_t box_size = bbox->dims()[1];

  std::vector<T> scores_data(num_boxes);
  std::copy_n(scores->data<T>(), num_boxes, scores_data.begin());
  std::vector<std::pair<T, int>> sorted_indices =
      GetSortedScoreIndex<T>(scores_data);

  std::vector<int> selected_indices;
  int selected_num = 0;
  T adaptive_threshold = nms_threshold;
  const T *bbox_data = bbox->data<T>();
  while (sorted_indices.size() != 0) {
    int idx = sorted_indices.back().second;
    bool flag = true;
    for (int kept_idx : selected_indices) {
      if (flag) {
        T overlap = JaccardOverlap<T>(bbox_data + idx * box_size,
                                      bbox_data + kept_idx * box_size, false);
        flag = (overlap <= adaptive_threshold);
      } else {
        break;
      }
    }
    if (flag) {
      selected_indices.push_back(idx);
      ++selected_num;
    }
    sorted_indices.erase(sorted_indices.end());
    if (flag && eta < 1 && adaptive_threshold > 0.5) {
      adaptive_threshold *= eta;
    }
  }
  return VectorToTensor(selected_indices, selected_num);
}

template <typename T>
class GenerateProposalsKernel : public framework::OpKernel<T> {
 public:
  void Compute(const framework::ExecutionContext &context) const override {
    auto *scores = context.Input<Tensor>("Scores");
    auto *bbox_deltas = context.Input<Tensor>("BboxDeltas");
    auto *im_info = context.Input<Tensor>("ImInfo");
    auto anchors = detail::Ref(context.Input<Tensor>("Anchors"),
                               "Cannot find input Anchors(%s) in scope",
                               context.Inputs("Anchors")[0]);
    auto variances = detail::Ref(context.Input<Tensor>("Variances"),
                                 "Cannot find input Variances(%s) in scope",
                                 context.Inputs("Variances")[0]);

    auto *rpn_rois = context.Output<LoDTensor>("RpnRois");
    auto *rpn_roi_probs = context.Output<LoDTensor>("RpnRoiProbs");

    int pre_nms_top_n = context.Attr<int>("pre_nms_topN");
    int post_nms_top_n = context.Attr<int>("post_nms_topN");
    float nms_thresh = context.Attr<float>("nms_thresh");
    float min_size = context.Attr<float>("min_size");
    float eta = context.Attr<float>("eta");

    auto &dev_ctx =
        context.template device_context<platform::CPUDeviceContext>();

    auto &scores_dim = scores->dims();
    int64_t num = scores_dim[0];
    int64_t c_score = scores_dim[1];
    int64_t h_score = scores_dim[2];
    int64_t w_score = scores_dim[3];

    auto &bbox_dim = bbox_deltas->dims();
    int64_t c_bbox = bbox_dim[1];
    int64_t h_bbox = bbox_dim[2];
    int64_t w_bbox = bbox_dim[3];

    rpn_rois->mutable_data<T>({bbox_deltas->numel() / 4, 4},
                              context.GetPlace());
    rpn_roi_probs->mutable_data<T>({scores->numel(), 1}, context.GetPlace());

    Tensor bbox_deltas_swap, scores_swap;
    bbox_deltas_swap.mutable_data<T>({num, h_bbox, w_bbox, c_bbox},
                                     dev_ctx.GetPlace());
    scores_swap.mutable_data<T>({num, h_score, w_score, c_score},
                                dev_ctx.GetPlace());

    math::Transpose<platform::CPUDeviceContext, T, 4> trans;
    std::vector<int> axis = {0, 2, 3, 1};
    trans(dev_ctx, *bbox_deltas, &bbox_deltas_swap, axis);
    trans(dev_ctx, *scores, &scores_swap, axis);

    framework::LoD lod;
    lod.resize(1);
    auto &lod0 = lod[0];
    lod0.push_back(0);
    anchors.Resize({anchors.numel() / 4, 4});
    variances.Resize({variances.numel() / 4, 4});

    int64_t num_proposals = 0;
    for (int64_t i = 0; i < num; ++i) {
      Tensor im_info_slice = im_info->Slice(i, i + 1);
      Tensor bbox_deltas_slice = bbox_deltas_swap.Slice(i, i + 1);
      Tensor scores_slice = scores_swap.Slice(i, i + 1);

      bbox_deltas_slice.Resize({h_bbox * w_bbox * c_bbox / 4, 4});
      scores_slice.Resize({h_score * w_score * c_score, 1});

      std::pair<Tensor, Tensor> tensor_pair =
          ProposalForOneImage(dev_ctx, im_info_slice, anchors, variances,
                              bbox_deltas_slice, scores_slice, pre_nms_top_n,
                              post_nms_top_n, nms_thresh, min_size, eta);
      Tensor &proposals = tensor_pair.first;
      Tensor &scores = tensor_pair.second;

      AppendProposals(rpn_rois, 4 * num_proposals, proposals);
      AppendProposals(rpn_roi_probs, num_proposals, scores);
      num_proposals += proposals.dims()[0];
      lod0.push_back(num_proposals);
    }
    rpn_rois->set_lod(lod);
    rpn_roi_probs->set_lod(lod);
    rpn_rois->Resize({num_proposals, 4});
    rpn_roi_probs->Resize({num_proposals, 1});
  }

  std::pair<Tensor, Tensor> ProposalForOneImage(
      const platform::CPUDeviceContext &ctx, const Tensor &im_info_slice,
      const Tensor &anchors, const Tensor &variances,
      const Tensor &bbox_deltas_slice,  // [M, 4]
      const Tensor &scores_slice,       // [N, 1]
      int pre_nms_top_n, int post_nms_top_n, float nms_thresh, float min_size,
      float eta) const {
    auto *scores_data = scores_slice.data<T>();

    // Sort index
    Tensor index_t;
    index_t.Resize({scores_slice.numel()});
    int *index = index_t.mutable_data<int>(ctx.GetPlace());
    for (int i = 0; i < scores_slice.numel(); ++i) {
      index[i] = i;
    }
    auto compare = [scores_data](const int64_t &i, const int64_t &j) {
      return scores_data[i] > scores_data[j];
    };

    if (pre_nms_top_n <= 0 || pre_nms_top_n >= scores_slice.numel()) {
      std::sort(index, index + scores_slice.numel(), compare);
    } else {
      std::nth_element(index, index + pre_nms_top_n,
                       index + scores_slice.numel(), compare);
      index_t.Resize({pre_nms_top_n});
    }

    Tensor scores_sel, bbox_sel, anchor_sel, var_sel;
    scores_sel.mutable_data<T>({index_t.numel(), 1}, ctx.GetPlace());
    bbox_sel.mutable_data<T>({index_t.numel(), 4}, ctx.GetPlace());
    anchor_sel.mutable_data<T>({index_t.numel(), 4}, ctx.GetPlace());
    var_sel.mutable_data<T>({index_t.numel(), 4}, ctx.GetPlace());

    CPUGather<T>(ctx, scores_slice, index_t, &scores_sel);
    CPUGather<T>(ctx, bbox_deltas_slice, index_t, &bbox_sel);
    CPUGather<T>(ctx, anchors, index_t, &anchor_sel);
    CPUGather<T>(ctx, variances, index_t, &var_sel);

    Tensor proposals;
    proposals.mutable_data<T>({index_t.numel(), 4}, ctx.GetPlace());
    BoxCoder<T>(ctx, &anchor_sel, &bbox_sel, &var_sel, &proposals);

    ClipTiledBoxes<T>(ctx, im_info_slice, &proposals);

    Tensor keep;
    FilterBoxes<T>(ctx, &proposals, min_size, im_info_slice, &keep);

    Tensor scores_filter;
    bbox_sel.mutable_data<T>({keep.numel(), 4}, ctx.GetPlace());
    scores_filter.mutable_data<T>({keep.numel(), 1}, ctx.GetPlace());
    CPUGather<T>(ctx, proposals, keep, &bbox_sel);
    CPUGather<T>(ctx, scores_sel, keep, &scores_filter);
    if (nms_thresh <= 0) {
      return std::make_pair(bbox_sel, scores_filter);
    }

    Tensor keep_nms = NMS<T>(ctx, &bbox_sel, &scores_filter, nms_thresh, eta);

    if (post_nms_top_n > 0 && post_nms_top_n < keep_nms.numel()) {
      keep_nms.Resize({post_nms_top_n});
    }

    proposals.mutable_data<T>({keep_nms.numel(), 4}, ctx.GetPlace());
    scores_sel.mutable_data<T>({keep_nms.numel(), 1}, ctx.GetPlace());
    CPUGather<T>(ctx, bbox_sel, keep_nms, &proposals);
    CPUGather<T>(ctx, scores_filter, keep_nms, &scores_sel);

    return std::make_pair(proposals, scores_sel);
  }
};

class GenerateProposalsOpMaker : public framework::OpProtoAndCheckerMaker {
 public:
  void Make() override {
    AddInput("Scores", "The scores of anchors should be foreground.");
    AddInput("BboxDeltas", "bbox_deltas.");
    AddInput("ImInfo", "Information for image reshape.");
    AddInput("Anchors", "All anchors.");
    AddInput("Variances", " variances");

    AddOutput("RpnRois", "Anchors.");
    AddOutput("RpnRoiProbs", "Anchors.");
    AddAttr<int>("pre_nms_topN", "pre_nms_topN");
    AddAttr<int>("post_nms_topN", "post_nms_topN");
    AddAttr<float>("nms_thresh", "nms_thres");
    AddAttr<float>("min_size", "min size");
    AddAttr<float>("eta", "The parameter for adaptive NMS.");
    AddComment(R"DOC(
Generate Proposals OP

This operator proposes rois according to each box with their probability to be a foreground object and 
the box can be calculated by anchors. Bbox_details and scores are the output of RPN. Final proposals
could be used to train detection net.

Scores is the probability for each box to be an object. In format of (N, A, H, W) where N is batch size, A is number
of anchors, H and W are height and width of the feature map.
BboxDeltas is the differece between predicted box location and anchor location. In format of (N, 4*A, H, W)

For generating proposals, this operator transposes and resizes scores and bbox_deltas in size of (H*W*A, 1) and (H*W*A, 4) and 
 calculate box locations as proposals candidates. Then clip boxes to image and remove predicted boxes with small area. 
Finally, apply nms to get final proposals as output.
)DOC");
  }
};

}  // namespace operators
}  // namespace paddle

namespace ops = paddle::operators;
REGISTER_OPERATOR(generate_proposals, ops::GenerateProposalsOp,
                  ops::GenerateProposalsOpMaker,
                  paddle::framework::EmptyGradOpMaker);
REGISTER_OP_CPU_KERNEL(generate_proposals, ops::GenerateProposalsKernel<float>,
                       ops::GenerateProposalsKernel<double>);