Generating binary tags for fast medical image retrieval based on convolutional nets and Radon Transform

Liu, Xinran; Tizhoosh, Hamid R.; Kofman, Jonathan

doi:10.1109/ijcnn.2016.7727562

Cited by 52 publications

(28 citation statements)

References 24 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Anavi et al (2016) and Liu et al (2016b) applied their methods to databases of X-ray images. Both used a five-layer CNN and extracted features from the fully-connected layers.…”

Section: Registrationmentioning

confidence: 99%

“…They showed that incorporating gender information resulted in better performance than just CNN features. Liu et al (2016b) used the penultimate fully-connected layer and a custom CNN trained to classify X-rays in 193 classes to obtain the descriptive feature vector. After descriptor binarization and data retrieval using Hamming separation values, the performance was inferior to the state of the art, which the authors attributed to small patch sizes of 96 pixels.…”

Section: Registrationmentioning

confidence: 99%

See 1 more Smart Citation

A survey on deep learning in medical image analysis

Litjens

Kooi

Bejnordi

et al. 2017

Medical Image Analysis

10,102

5,414

View full text Add to dashboard Cite

Deep learning algorithms, in particular convolutional networks, have rapidly become a methodology of choice for analyzing medical images. This paper reviews the major deep learning concepts pertinent to medical image analysis and summarizes over 300 contributions to the field, most of which appeared in the last year. We survey the use of deep learning for image classification, object detection, segmentation, registration, and other tasks. Concise overviews are provided of studies per application area: neuro, retinal, pulmonary, digital pathology, breast, cardiac, abdominal, musculoskeletal. We end with a summary of the current state-of-the-art, a critical discussion of open challenges and directions for future research.

show abstract

“…Anavi et al (2016) and Liu et al (2016b) applied their methods to databases of X-ray images. Both used a five-layer CNN and extracted features from the fully-connected layers.…”

Section: Registrationmentioning

confidence: 99%

Section: Registrationmentioning

confidence: 99%

A survey on deep learning in medical image analysis

Litjens

Kooi

Bejnordi

et al. 2017

Medical Image Analysis

10,102

5,414

View full text Add to dashboard Cite

show abstract

“…10 Common data standardization methods include simple scaling, unit length scaling, sample-by-sample mean subtraction, and feature standardization. Standardization is a kind of "minification" conversion of data eigenvalues, and standardized data are scaled to a more reasonable range.…”

Section: Multi-feature Ct Image Preprocessingmentioning

confidence: 99%

“…Standardization is a kind of "minification" conversion of data eigenvalues, and standardized data are scaled to a more reasonable range. 10 Common data standardization methods include simple scaling, unit length scaling, sample-by-sample mean subtraction, and feature standardization. 1 Simple scaling Simple scaling has important implications for data processing.…”

Section: Multi-feature Ct Image Preprocessingmentioning

confidence: 99%

Multi‐feature data mining for CT image recognition

Yong-lian

2018

Concurrency and Computation

View full text Add to dashboard Cite

Recognition method of CT images based on color, morphology, and texture is inaccurate and unreliable recently. Therefore, a medical image recognition method based on data mining is proposed with a multi-feature fusion in this paper. First, image acquisition method is determined by analyzing the environment of CT image recognition. Then, acquired CT image is standardized and whitened to reduce redundant information. Moreover, based on feature of the color, texture, and height in preprocessed CT images, a deep neural network is trained by using a large amount of image data in normal scene. The deep learning classifier is fine-tuned by using the marked multi-feature CT image data. Finally, output recognition result is obtained according to the classification decision threshold. Experimental results show that the correct recognition rate of the proposed method can reach more than 98%. The accuracy rate is higher and the stability of proposed method is better by comparing with traditional CT image recognition methods. KEYWORDSCT images, data mining, deep learning, image recognition, multiple feature INTRODUCTIONIn recent years, with the rapid development of medical imaging technology, it improved to big medical image analysis. Today, many big database are constructed. How we can extract useful information from massive medical image data in the database has brought great challenges to medical image recognition since traditional database technology cannot provide enough service for the large database. Data mining is used to obtain effective information in medical image database. CT image recognition is a multidisciplinary cross-domain of comprehensive medical imaging, mathematical modeling, and computer technology. 1,2 Today, there are two new issues that are brought because of large scale of multi-feature CT image. One issue is CT image data that has a higher dimension and requires a model with stronger learning adaptability. The other is big CT image data that is more fragmented with more complex data structure, which requires integration of huge different information. 3 Such requirements cannot be solved by traditional data analysis methods. Therefore, how we extract useful information from massive CT image data has become a research hotspot. With the rapid development and popularization of CT imaging equipment, CT image data have increased explosively. How we can efficiently and accurately analyze CT image has become a major challenge. 4 Furthermore, computer-assisted CT medical image analysis improves the efficiency and accuracy of the analysis and diagnosis of CT image. Therefore, computer-aided detection and diagnosis are becoming a cross-sectional research area that receives more and more attention.A multi-feature CT medical image recognition method was proposed based on polarization spectrum. 5 The method used indoor imaging spectrometer FISS-P equipped with polarizers to collect the polarization spectrum images, including general images and five kinds of CT medical images indoors. They recognized CT images by ...

show abstract

“…Deep denoising autoencoder (DDA), for instance, was used to hash the X-ray images into binary codes [15]. Deep convolutional based image retrieval neural networks have also been investigated [16], [17]. The drawback of such solutions is that they require a large, labeled and balanced dataset, and a lot of computational resources for training.…”

Section: Image Retrievalmentioning

confidence: 99%

Combining real-valued and binary Gabor-Radon features for classification and search in medical imaging archives

Erfankhah

Yazdi

Tizhoosh

2017

2017 IEEE Symposium Series on Computational Intelligence (SSCI)

View full text Add to dashboard Cite

Content-based image retrieval (CBIR) of medical images in large datasets to identify similar images when a query image is given can be very useful in improving the diagnostic decision of the clinical experts and as well in educational scenarios. In this paper, we used two stage classification and retrieval approach to retrieve similar images. First, the Gabor filters are applied to Radon-transformed images to extract features and to train a multi-class SVM. Then based on the classification results and using an extracted Gabor barcode, similar images are retrieved. The proposed method was tested on IRMA dataset which contains more than 14,000 images. Experimental results show the efficiency of our approach in retrieving similar images compared to other Gabor-Radon-oriented methods.

show abstract

Generating binary tags for fast medical image retrieval based on convolutional nets and Radon Transform

Cited by 52 publications

References 24 publications

A survey on deep learning in medical image analysis

A survey on deep learning in medical image analysis

Multi‐feature data mining for CT image recognition

Combining real-valued and binary Gabor-Radon features for classification and search in medical imaging archives

Contact Info

Product

Resources

About