Existing image captioning models are usually trained by cross-entropy (XE) loss and reinforcement learning (RL), which set ground-truth words as hard targets force the model to learn from them. However, widely adopted training strategies may suffer misalignment in XE inappropriate reward assignment RL training. To tackle these problems, we introduce an attribute enhanced teacher that serves a b...

this paper presents a comparison study between the multilayer perceptron (mlp) and radial basis function (rbf) neural networks with supervised learning and back propagation algorithm to track hand gestures. both networks have two output classes which are hand and face. skin is detected by a regional based algorithm in the image, and then networks are applied on video sequences frame by frame in...

Self-training is a simple semi-supervised learning approach: Unlabelled examples that attract high-confidence predictions are labelled with their and added to the training set, this process being repeated multiple times. Recently, self-supervision—learning without manual supervision by solving an automatically-generated pretext task—has gained prominence in deep learning. This paper investigate...

B-mode ultrasound imaging is a popular medical technique. Like other image processing tasks, deep learning has been used for analysis of images in the last few years. However, training models require large labeled datasets, which often unavailable images. The lack data bottleneck use analysis. To overcome this challenge, work, we exploit auxiliary classifier generative adversarial network (ACGA...

The Image Difference Captioning (IDC) task aims to describe the visual differences between two similar images with natural language. major challenges of this lie in aspects: 1) fine-grained that require learning stronger vision and language association 2) high-cost manual annotations leads limited supervised data. To address these challenges, we propose a new modeling framework following pre-tr...

We propose a novel semi-supervised learning approach for classification of histopathology images. employ strong supervision with patch-level annotations combined co-training loss to create framework. Co-training relies on multiple conditionally independent and sufficient views the data. separate hematoxylin eosin channels in pathology images using color deconvolution two each slide that can par...