HyperNet: Towards Accurate Region Proposal Generation and Joint Object Detection

Abstract

Almost all of the current top-performing object detection networks employ region proposals to guide the search for object instances. State-of-the-art region proposal methods usually need several thousand proposals to get high recall, thus hurting the detection efficiency. Although the latest Region Proposal Network method gets promising detection accuracy with several hundred proposals, it still struggles in small-size object detection and precise localization (e.g., large IoU thresholds), mainly due to the coarseness of its feature maps. In this paper, we present a deep hierarchical network, namely HyperNet, for handling region proposal generation and object detection jointly. Our HyperNet is primarily based on an elaborately designed Hyper Feature which aggregates hierarchical feature maps first and then compresses them into a uniform space. The Hyper Features well incorporate deep but highly semantic, intermediate but really complementary, and shallow but naturally high-resolution features of the image, thus enabling us to construct HyperNet by sharing them both in generating proposals and detecting objects via an end-to-end joint training strategy. For the deep VGG16 model, our method achieves completely leading recall and state-of-the-art object detection accuracy on PASCAL VOC 2007 and 2012 using only 100 proposals per image. It runs with a speed of 5 fps (including all steps) on a GPU, thus having the potential for real-time processing.

A hybrid framework combining background subtraction and deep neural networks for rapid person detection

“…Most of the recent object detection methods are based on deep learning adopting CNNs [41][42][43]. Since the advent of the R-CNN [44], combining region proposal and CNN classification became a the preferred framework for object detection.…”

Aircraft detection in remote sensing images based on saliency and convolution neural network

“…For example, many researchers regard that the lower layer of the neural network usually retain more fine-grained, while the higher layer of it usually has better semantic features. To improve the accuracy of detection, the features of different layer are fused [59] [69][70][71][72][73]. Other researchers combine the object detection with other applications of image processing, e.g., Kaiming [74] combine the object detection with image segmentation and gets good results.…”

“…These algorithms are mainly classified into two groups: one is object detection method based on region proposal [58][59][60], which is a mainstream algorithm, e.g., RCNN [31], SPPNet [61], Fast-RCNN [34], Faster-RCNN [62], and MSRA recently proposes algorithm R-FCN [63]. The other is not using the region proposal method to detection, e.g., YOLO [64] and SSD [65].…”

Automatic Raft Labeling for Remote Sensing Images via Dual-Scale Homogeneous Convolutional Neural Network

“…Early fusion is feature-level, which refers to merging kinds of features into stronger ones. There are two classical early fusion methods: high-level and low-level fusion by skip connections [37,38] and multi-scale fusion by constructing feature pyramid [20,24,39].…”