Unlock instant, AI-driven research and patent intelligence for your innovation.

Fish visual identification method based on multi-task fusion

A technology of fish vision and recognition method, applied in the field of target detection, can solve problems such as inability to perform parallel processing, and achieve the effect of improving the speed of reasoning

Pending Publication Date: 2022-08-02
DALIAN OCEAN UNIV
View PDF0 Cites 0 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

[0004] At present, few scholars have conducted research on multi-task networks in the field of fishery, but in the field of autonomous driving, there have been a large number of applications using multi-task networks in panoramic perception [Teichmann, Marvin, et al. "Multinet: Real-time joint semantic reasoning for autonomous driving."2018IEEE Intelligent Vehicles Symposium(IV).IEEE,2018.】
MultiNet uses one encoder and three decoder branches to achieve scene classification, target detection and semantic segmentation of driving areas, and achieves real-time detection, but this architecture is only applicable to the field of autonomous driving
LSNet [Duan, Kaiwen, et al. "Location-sensitive visual recognition with cross-iou loss." arXiv preprint arXiv:2104.04899(2021).] achieves target detection with a unified framework by uniformly encoding location-sensitive tasks , instance segmentation, pose estimation, but cannot perform three tasks in parallel

Method used

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
View more

Image

Smart Image Click on the blue labels to locate them in the text.
Viewing Examples
Smart Image
  • Fish visual identification method based on multi-task fusion
  • Fish visual identification method based on multi-task fusion
  • Fish visual identification method based on multi-task fusion

Examples

Experimental program
Comparison scheme
Effect test

Embodiment

[0042] In this example, 2.6k fish images are collected and manually labeled. The labeling results follow the MS-COCO format. The labeling content includes the real target frame, instance segmentation mask polygons, and pose estimation key point coordinates.

[0043] This example trains the framework of this paper on NVIDIA TESLA v100. This example tries different combinations of hyperparameters and optimizers to find the most suitable training method. In this example, the input image is uniformly resized to 416×416 for inference. This example uses the pre-trained model of Darknet53 on ImageNet [imagenet] as the initialization weight. In this embodiment, the average precision (AP) is used to measure the model performance of this embodiment, and the relevant calculation standard is consistent with MS-COCO. For object detection, this embodiment uses the intersection of prediction frame and ground truth (IOU) threshold (from 0.5 to 0.95) to calculate, for instance segmentation, th...

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to View More

PUM

No PUM Login to View More

Abstract

According to the fish visual identification method based on multi-task fusion, target detection, instance segmentation and attitude estimation can be carried out in parallel by using an efficient multi-task network, and the reasoning speed is increased to a real-time level. Compared with a base line, the network provided by the invention only adds negligible parameter quantity, and the processing capability of the network on parallel multiple tasks is mined, so that the network provided by the invention can be easily deployed in an actual aquaculture application scene. According to the invention, the thought of predicting a multi-task network by using a single decoder branch is provided, and heuristic is provided for fusion coding among a plurality of tasks.

Description

technical field [0001] The invention belongs to the field of target detection, in particular to a fish visual recognition method based on multi-task fusion. Background technique [0002] In the process of aquaculture, the detection of the physiological state and movement behavior of fish can better realize the precise breeding process. The acquisition of physiological state and movement behavior can be carried out by calculating the body length, weight and movement posture of fish. Object detection, keypoint detection, and instance segmentation in the field of computer vision can provide accurate prediction of target frame, mask and keypoint skeleton information. The predicted target frame obtained by target detection can accurately locate the position of the fish, the mask obtained by instance segmentation can obtain the outline of the fish, and the key point information can judge the movement and posture of the fish. [0003] With the development of deep learning, there ...

Claims

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to View More

Application Information

Patent Timeline
no application Login to View More
Patent Type & Authority Applications(China)
IPC IPC(8): G06V10/46G06V10/764G06K9/62G06V10/82G06N3/04G06N3/08
CPCG06N3/08G06N3/045G06F18/241Y02A40/81
Inventor 曹立杰陈子文王其华
Owner DALIAN OCEAN UNIV