Camouflaged objects detection network via contradiction detection and feature aggregation

A Linear Differentiation Scheme for Camouflaged Target Detection using Convolution Neural Networks

“…Bi et al [14] proposed a COD framework named CFNet, CFNet, which is mainly used to solve problems in the detection process. The actual goal of the framework is to reduce the blurred edges and difficulties in object detection.…”

“…The boundary-based approach in [16] requires certain textural classifications regardless of the image size and pixel representations. The contradiction method discussed in [14] requires deviating instances that segregate the pixels, resulting in false negatives. Such issues are addressed in this proposed scheme for improving the positive rate of object detection.…”