Background Mixup Data Augmentation for Hand and Object-in-Contact Detection

Abstract

Detecting the positions of human hands and objects-incontact (hand-object detection) in each video frame is vital for understanding human activities from videos. For training an object detector, a method called Mixup, which overlays two training images to mitigate data bias, has been empirically shown to be effective for data augmentation. However, in hand-object detection, mixing two hand-manipulation images produces unintended biases, e.g., the concentration of hands and objects in a specific region degrades the ability of the hand-object detector to identify object boundaries. We propose a data-augmentation method called Background Mixup that leverages data-mixing regularization while reducing the unintended effects in hand-object detection. Instead of mixing two images where a hand and an object in contact appear, we mix a target training image with background images without hands and objects-in-contact extracted from external image sources, and use the mixed images for training the detector. Our experiments demonstrated that the proposed method can effectively reduce false positives and improve the performance of hand-object detection in both supervised and semi-supervised learning settings.

ICIP 2022 Challenge on Parasitic Egg Detection and Classification in Microscopic Images: Dataset, Methods and Results

“…All methods in this competition exploited deep learning techniques. Most of them were developed based on the state of the art in object detection, including YOLOv5 [15,16], Fast-RCNN [13,17], EfficientDet [18], Cascade R-CNN [15,19,20], CBNetV2 [21,22], CenterNet2 [23], Task-aligned One-stage object Detection (TOOD) [17,24], and RetinaNet [25,26]. These methods are convolutional neural networks with various backbone architectures, where the most popular architecture is based on ResNet blocks [17, 22-25, 27, 28].…”