Patent
Automatic squid classification method based on color image and convolutional neural network technology
Hu Jun,Chen Wenxuan,Chengquan Zhou,Zhao Dandan +3 more
- 01 Oct 2019
TL;DR: Wang et al. as mentioned in this paper presented an automatic squid classification method based on a color image and a convolutional neural network technology, which comprises the following steps: unfreezing squids, cleaning the unfrozen squids and removing pollutants on the surfaces of the squids.
read more
Abstract: The invention discloses an automatic squid classification method based on a color image and a convolutional neural network technology, which comprises the following steps: unfreezing squids, cleaningthe unfrozen squids, removing pollutants on the surfaces of the squids, and removing damaged squids to prepare squid samples for automatic classification of the squids; flatly unfolding the squid sample on work, placing the squid sample in an auxiliary light source irradiation area, and performing image acquisition on the squid sample at different angles by using shooting equipment to obtain an original squid image; carrying out image preprocessing on the original squid image to obtain a test image; inputting the test image into an improved convolutional neural network for training to obtain image features of different types of squids so as to realize squid classification and recognition. According to the method, the machine vision technology and the deep learning framework are combined, the features of the squids in the image are automatically extracted through the improved convolutional neural network, correct classification of different varieties is achieved, and therefore the requirement for fine processing in industrialized production is met.
read more
Chat with Paper
AI Agents for this Paper
Find similar papers on Google Scholar, PubMed and Arxiv
Write a critical review of this paper
Analyze citations of this paper to find unaddressed research gaps
References
Patent
Image classification method based on convolution neural network
Wang Lei,He Yueshun,Wang Kun,Jiang Niande,Zhong Guoyun,Cai Youlin +5 more
- 10 Nov 2017
TL;DR: Zhang et al. as mentioned in this paper revealed an image classification method based on a convolution neural network, which comprises the following steps: constructing a deep CNN, improving the CNN, training and testing, and optimizing the CNN parameters.
54
Patent
Image processing-based target fish recognition method and system
Duan Qingling,Li Daoliang,Hong Yao,Sun Longqing,Chen Yingyi +4 more
- 20 Nov 2013
TL;DR: In this paper, an image processing-based target fish recognition method was proposed, which comprises the following steps: acquiring a fish body color image; performing image segmentation on the fish body colour image to acquire a target image; extracting the color feature, the texture feature and the shape feature of the target image to obtain comprehensive feature vectors; and performing classified processing on a test set combined by the comprehensive feature vector by using an RBF (Radial Basis Function) neural network to recognize whether a fish is a target fish.
17
Patent
Deep learning-based fish positioning detection and recognition method and system
Yingyi Chen,Chuanyang Gong,Yeqi Liu,Xiaomin Fang,Cheng Qianqian,Cheng Yanjun,Yu Huihui +6 more
- 13 Nov 2018
TL;DR: In this paper, a deep learning-based fish positioning detection and recognition method and system is proposed, which comprises the steps of: inputting a to-be-recognized image into a first preset neural network and obtaining target frames corresponding to fishes in the to be recognized image according to output results of the first neural network; and inputting images corresponding to the target frames into a second-preview neural network, and obtaining species of the fishes in target frames according to the output of the second neural network.
9
Patent
Data processing system with means for contiguously addressing memory
Dag Reidar Blokkum,Charles Ray Johns,Lee Jack Morozink,David Lawrence Peterson +3 more
- 29 Mar 1988
TL;DR: In this article, the authors propose a method for contiguously addressing a memory system comprising a plurality of memory banks, each of which can have different capacity modules, or no module, plugged therein, comprises a matrix of logic cells.
8