Detection and Recognition of Bangladeshi Fishes Using Surf and Convolutional Neural Network.

Abstract

This paper represents a model to detection and recognize local fishes of Bangladesh implementing image processing and neural networking approaches. The aim of the research work is to apply computer vision and AI techniques so that people of next generation can recognize Bangladeshi fishes as most of the young people in city, have less idea to classify traditional and deshi fishes. We implemented our custom Dataset consisting of 400 sample images for the experiment method to measure out its credibility. In the proposed, model sequential grassfire algorithm is used along with pre-processing techniques like noise cancelation, gray scaling, flood-fill method, binarization to detect and analysis shape of fish. Further, to do classification and recognition of the detected fishes, convolutional neural network (CNN) and method of Speeded up robust feature (SURF) had been applied to visualize difference between to techniques. CNN architecture got better accuracy with 90.9% score for recognition and classify fishes where SURF algorithm visualize better recognition matching putative points after extracting features.

Fish Species Identification Techniques: A Review

“…So, there was no need to use image prefiltering. The results show that the use of multi-level residual network [25], machine learning algorithm [17], SURF [11] achieved high accuracy with 99.69%, 98% and 98.67% respectively. Maximum number of classified fish species was 38 species [17].…”

“…Combination of image processing techniques such as image enhancement and segmentation can used to improve fish classification system [9], images pre-processing such as rotation, reflection, histogram, equalization, translation and gaussian blurring, data enhancement process was performed on distorted copies of existing images which led to improvement of the work of the system through expand the data set [10]. A gain gaussian blur method was used to blur input image, it used to reduce noise followed by such operations: binary masking of image, flood fill operation (sometimes known as a seed fill), boundary extraction using the sequential grass-fire algorithm [11]. Cropping fish image, edge detection and cut out 512512-pixel size sub-image was used in [12].…”

“…Detect and classify fish [13] Convolutional neural networks Not reported [27] Convolutional neural networks 90% and 92% for detection and classification Fish detection / recognition [11] AI techniques and computer vision 90.9% for CNN, 87.08 -98.67% for SURF…”