TransCenter: Transformer in Heatmap and a New Form of Bounding Box

EasyChair Preprint 11138

Abstract

In current heatmap-based object detection, the task of heatmap is to predict the position of keypoints and its category. However, since objects of the same category share the same channel in the heatmap, it is possible for their keypoints to overlap. When this phenomenon occurs, existing heatmap-based detectors are unable to differentiate between the overlapping keypoints. To address the above is-sue, we have designed a new heatmap-based object detection model, called TransCenter. Our model decouples the tasks of predicting the object category and keypoint position, and treats object detection as a set prediction task. We use a label assignment strategy to divide the predicted sets into positive and negative samples for training. The purpose of this is to allow different objects to have their own heatmap channel without sharing with other, thereby completely eliminating the occurrence of overlapping. To make the model easier to learn, we leverage the characteristic that heatmaps can reduce the solution space, proposed a novel approach for predicting bounding boxes. We use the encoder-decoder structure in transformers, treat the prediction of bounding boxes as an encoding task, use the form of a heatmap to represent the position and size. Then, we treat category prediction and offset prediction of the bounding box as decoding tasks, where the offset prediction is outputted through regression.

Keyphrases: Keypoint, Keypoint Overlap, heatmap, object detection

@booklet{EasyChair:11138,
  author    = {Deqi Liu and Aimin Li and Mengfan Cheng and Dexu Yao},
  title     = {TransCenter: Transformer in Heatmap and a New Form of Bounding Box},
  howpublished = {EasyChair Preprint 11138},
  year      = {EasyChair, 2023}}