Survey on deep learning for radiotherapy.

The content based image retrieval aims to find the similar images from a large scale dataset against a query image. Generally, the similarity between the representative features of the query image and dataset images is used to rank the images for retrieval. In early days, various hand designed feature descriptors have been investigated based on the visual cues such as color, texture, shape, etc. that represent the images. However, the deep learning has emerged as a dominating alternative of hand-designed feature engineering from a decade. It learns the features automatically from the data. This paper presents a comprehensive survey of deep learning based developments in the past decade for content based image retrieval. The categorization of existing state-of-the-art methods from different perspectives is also performed for greater understanding of the progress. The taxonomy used in this survey covers different supervision, different networks, different descriptor type and different retrieval type. A performance analysis is also performed using the state-of-the-art methods. The insights are also presented for the benefit of the researchers to observe the progress and to make the best choices. The survey presented in this paper will help in further research progress in image retrieval using deep learning.

A Decade Survey of Content Based Image Retrieval using Deep Learning

Optimization of deep convolutional neural network for large scale image retrieval

Content based image retrieval with sparse representations and local feature descriptors : A comparative study

Image retrieval based on query by saliency content

To build an industrial content-based image retrieval system (CBIRs), it is highly recommended that feature extraction, feature processing and feature indexing need to be fully considered. Although research that bloomed in the past years suggest that the convolutional neural network (CNN) be in a leading position on feature extraction & representation for CBIRs, there are less instructions on the deep analysis of feature related topics, for example the kind of feature representation that has the best performance among the candidates provided by CNN, the extracted features generalization ability, the relationship between the dimensional reduction and the accuracy loss in CBIRs, the best distance measure technique in CBIRs and the benefit of the coding techniques in improving the efficiency of CBIRs, etc. Therefore, several practicing studies were conducted and a thorough analysis was made in this research attempting to answer the above questions. The results in the study on both ImageNet-2012 and an industrial dataset provided by Sogou demonstrate that fc4096a and fc4096b perform the best on the datasets from unseen categories. Several interesting and practicing conclusions are drawn, for instance, fc4096a and fc4096b are found to have a better generalization ability than other features of CNN and could be considered as the first choice for industrial CBIRs. Furthermore, a novel feature binarization approach is presented in this paper for better efficiency of CBIRs. More specifically, the binarization is capable of reducing 31/32 space usage of original data. To sum up, the conclusions seem to provide practical instructions on real industrial CBIRs.

Deep Learning for Image Retrieval: What Works and What Doesn't

Image Stitching refers to the technology fusing more than one images with overlapping part into a large field of view image. Image mosaic consists of image preprocessing, image registration and image fusion. To solve problems of serious clustering phenomenon and fewer corner points in the texture region caused by traditional Harris Corner detection algorithm, this paper proposes an improving adaptive threshold setting algorithm by calculating the second-order value of the corner response function, avoiding effects of the selection of scale factor k and threshold T on corner detection. To overcome the weakness of obvious traces in the jointing places caused by traditional weighted average method for image fusion, this paper enhances the weighted average method with trigonometric functions. Experimental results show our proposed algorithms can effectively eliminate the gap generated by image mosaic, with a better speed and precision.

Yanxiang Zhang

Papers

Deep Learning for Image Retrieval: What Works and What Doesn't

An Adaptive Harris Corner Detection Algorithm for Image Mosaic