Posture detection targeted towards providing assessments for the monitoring of health and welfare of pigs has been of great interest to researchers from different disciplines. Existing studies applying machine vision techniques are mostly based on methods using three-dimensional imaging systems, or two-dimensional systems with the limitation of monitoring under controlled conditions. Thus, the main goal of this study was to determine whether a two-dimensional imaging system, along with deep learning approaches, could be utilized to detect the standing and lying (belly and side) postures of pigs under commercial farm conditions. Three deep learning-based detector methods, including faster regions with convolutional neural network features (Faster R-CNN), single shot multibox detector (SSD) and region-based fully convolutional network (R-FCN), combined with Inception V2, Residual Network (ResNet) and Inception ResNet V2 feature extractions of RGB images were proposed. Data from different commercial farms were used for training and validation of the proposed models. The experimental results demonstrated that the R-FCN ResNet101 method was able to detect lying and standing postures with higher average precision (AP) of 0.93, 0.95 and 0.92 for standing, lying on side and lying on belly postures, respectively and mean average precision (mAP) of more than 0.93.

https://www.mdpi.com/1424-8220/19/17/3738/pdf

Deep Learning and Machine Vision Approaches for Posture Detection of Individual Pigs

Cascading handcrafted features and Convolutional Neural Network for IoT-enabled brain tumor segmentation

Cancer disease was one of the significant health problems in the world that cause deaths. Skin cancer is a type of cancer caused by damaged DNA and the body cannot recover the damage. This makes the cells begin to grow and divide uncontrollably. Skin cancer in Indonesia was ranks third after breast and neck cancer. Cases of skin cancer reach 5.9-7.8% of all types of cancer every year. This disease could be identified using the dermatoscopic image. Through those images, doctor decided the skin condition based on the characteristics that exist. Decision making in determining the skin condition depends on the ability and knowledge of the doctor. Therefore, a skin cancer identification system is needed to help make decisions in identifying skin cancer. Convolutional neural network (CNN) was used in this identification system because it can be detecting image and pattern. CNN work through three stages, that is convolutional layer, pooling layer, and fully-connected layer. This identification system is based on the dermoscopy image of HAM10000 skin cancer dataset. Based on this research, the accuracy of training and testing of skins cancer identification system are 80% and 78%.

Skins cancer identification system of HAMl0000 skin cancer dataset using convolutional neural network

3D shape retrieval has attracted much research attention due to its wide applications in the fields of computer vision and multimedia. Various approaches have been proposed in recent years for learning 3D shape descriptor from different modalities. The existing works contain the following disadvantages: 1) the vast majority methods rely on the large scale of training data with clear category information; 2) many approaches focus on the fusion of multi-modal information but ignore the guidance of correlations among different modalities for shape representation learning; 3) many methods pay attention to the structural feature learning of 3D shape but ignore the guidance of structural similarity between every two shapes. To solve these problems, we propose a novel multi-graph network (MGN) for unsupervised 3D shape retrieval, which utilizes the correlations among modalities and structural similarity between two models to guide the shape representation learning process without category information. More specifically, we propose two novel loss functions: auto-correlation loss and cross-correlation loss. The auto-correlation loss utilizes information from different modalities to increase the discrimination of shape descriptor. The cross-correlation loss utilizes the structural similarity between two models to strengthen the intra-class similarity and increase the inter-class distinction. Finally, an effective similarity measurement is designed for the shape retrieval task. To validate the effectiveness of our proposed method, we conduct experiments on the ModelNet dataset. Experimental results demonstrate the effectiveness of our proposed method, and significant improvements have been achieved compared with state-of-the-art methods.

Multi-graph Convolutional Network for Unsupervised 3D Shape Retrieval

Nowadays, object detection has acquired a great concentration either in ordinary images or satellite images. For satellite images, object detection is a challenging problem because objects have different scales and sparsity with very complicated background. Recent deep learning approaches have achieved breaking results for object detection than traditional ones. The ability of bounding boxes to catch existing objects with a complete and precise manner is still a challenging problem. We propose a dilated anchor method based on You Only Look Once version 3(YOLOv3) algorithm to make object detection more flexible and precise. The proposed method uses greater size anchor bounding boxes with about 30 % to 40 % larger than the traditional ones. This increase in anchor size increases the ability to catch more class objects with less influence on location detection. The experimental results using public NWPU VHR-10 dataset demonstrate the effectiveness of the proposed method in object detection of most classes and increase the overall accuracy with minimal effect on the precise location.

Convolutional Neural Network with Dilated Anchors for Object Detection in Very High Resolution Satellite Images

Deep convolution neural network with scene-centric and object-centric information for object detection

In this paper, a novel framework to retrieve 3D articulated models from a database based on one or few depth images is presented. Existing state-of-the-arts retrieval approaches usually constrain the view points of query images or assume that the target models are rigid-body. When they are applied to retrieving articulated models, the retrieved results are substantially influenced by the model postures. In our work, we extracts the limbs and torso regions from projections and analyzes the features of local regions. The use of both global and local features can alleviate the disturbance of model postures in model retrieval. Experiments show that the proposed method can efficiently retrieve relevant models within a second, and provides higher retrieval accuracy than those of compared methods for not only rigid body 3D models but also models with articulated limbs.

Yo-Chung Lau

Papers

Deep convolution neural network with scene-centric and object-centric information for object detection

3D Articulated Model Retrieval Using Depth Image Input