Browsing by Author "Wang, Yi"

Now showing 1 - 2 of 2
  • Thumbnail Image
    Item
    Salient Object Detection for complex scenes : a thesis presented in partial fulfilment of the requirements for the degree of Doctor of Philosophy in Computer Science, School of Mathematical and Computational Sciences, Massey University, Albany, Auckland, New Zealand
    (Massey University, 2024-09-19) Wang, Yi
    Salient Object Detection (SOD), a primary objective in computer vision, aims to locate and segment the region most visually striking within an image. In this thesis, we present three innovative methods based on deep learning to improve SOD performance in complex scenes. Firstly, we introduce the Multiple Enhancement Network (MENet), inspired by boundary perception, gradual enhancement, frequency decomposition, and content integrity of the Human Visual System (HVS). We propose a flexible multi-scale feature enhancement module to aggregate and refine features and use iterative training to improve boundary and adaptive features in the dual-branch decoder of MENet. A multi-level hybrid loss guides the network in learning pixel-, region-, and object-level features. Evaluations of benchmark datasets show that MENet outperforms other SOD models, especially when the salient region has multiple objects with varied appearances or complex shapes. Secondly, we propose TFGNet, an effective frequency-guided network for saliency detection based on Transformer. TFGNet has a parallel two-branch decoder, which leverages a pixel-wise decoder and a Transformer decoder to optimise high-spatial frequency boundary details and low-spatial frequency salient features. A novel loss is also designed to use frequency distribution similarity measurement to further improve performance. The experimental results indicate that TFGNet can accurately locate salient objects with more complete and precise boundaries on various complex backgrounds. This framework also rekindles awareness of the advantages of exploiting images' spatial frequency features in SOD. Thirdly, we design a multi-source weakly supervised SOD (WSOD) framework that can effectively utilise pseudo-background (non-salient region) labels combined with scribble labels to obtain more accurate salient features. We first create a comprehensive salient pseudo-mask generator from multiple self-learning features. Also, we pioneer the exploration of generating salient pseudo-labels via point-prompted and box-prompted Segment-Anything Models (SAM). Then, a Transformer-based WSOD network named WBNet is proposed, which leverages pixel-decoder and transformer-decoder with auxiliary edge predictor with multi-source loss function to handle complex saliency detection tasks. In summary, we contribute three novel approaches to address salient object detection in complex scenes. Each model achieves cutting-edge performance across prestigious datasets validated through comprehensive experiments.
  • Thumbnail Image
    Item
    A web-based teleoperative mobile robotic system : Master of Engineering in Information Engineering at Massey University, Albany, Auckland, New Zealand
    (Massey University, 2006) Wang, Yi
    With the rapid development of internet technology, it becomes real that human beings can access, modify and control a remote hardware device via internet connection. Such remote operations can replace the human to be present at a dangerous or unreachable place or can make as many as possible users to access the hardware in different places at a low cost. The thesis research was aimed at developing a web based mobile robot control framework for education purpose. It should be composed of a mobile robot. Http server, dynamic user interface and video server. With it users can view and control the real robot via a normal web browser and can choose to run either simulation or the real robot. This is done by setting up operational parameters via a friendly GUI (graphic user interface). Users also can upload and compile their own C code to control the robot and get back the running results. The main objectives of this thesis research are hardware upgrading for Nomadic Super Scout mobile robot and web based php programming. For the first objective, the onboard PC was replaced by a laptop that is remotely placed and connected to the robot control system via Bluetooth wireless. The Nserver for robot simulation was set up in the Linux operating environment. For the second objective, the software programming was focused on building a web control platform which should be user friendly. An Apache server was developed where PHP program was used for the user interface. The main advantage of using PHP is that it does not need to install or download any software or script to get access to the remote robot via a normal web browser on any operation like windows or Linux. The web-based mobile robot system was tested using two different cases. One case demonstrated how the user specifies a set of motion parameters of the robot that is programmed to perform a wall-following behaviour. The other demonstrated how the user uploads a collision avoidance program to run the robot that is placed among obstacles. Both case studies were performed in real environments and the results proved the success of the developed web-based robotic system.