Exemplar-Based Recursive Instance Segmentation With Application to Plant Image Analysis

Deep Learning-Based Occlusion Handling of Overlapped Plants for Robotic Grasping

Instance segmentation of overlapping plants to detect their grasps for possible robotic grasping presents a challenging task due to the need to address the problem of occlusion. We addressed the problem of occlusion using a powerful convolutional neural network for segmenting objects with complex forms and occlusions. The network was trained with a novel dataset named the “occluded plants” dataset, containing real and synthetic images of plant cuttings on flat surfaces with differing degrees of occlusion. The synthetic images were created using the novel framework for synthesizing 2D images by using all plant cutting instances of available real images. In addition to the method for occlusion handling for overlapped plants, we present a novel method for determining the grasps of segmented plant cuttings that is based on conventional image processing. The result of the employed instance segmentation network on our plant dataset shows that it can accurately segment the overlapped plants, and it has a robust performance for different levels of occlusions. The presented plants’ grasp detection method achieved 94% on the rectangle metric which had an angular deviation of 30 degrees and an IoU of 0.50. The achieved results show the viability of our approach on plant species with an irregular shape and provide confidence that the presented method can provide a basis for various applications in the food and agricultural industries.

CycleSegNet: Object Co-Segmentation With Cycle Refinement and Region Correspondence

Image co-segmentation is an active computer vision task that aims to segment the common objects from a set of images. Recently, researchers design various learning-based algorithms to undertake the co-segmentation task. The main difficulty in this task is how to effectively transfer information between images to make conditional predictions. In this paper, we present CycleSegNet, a novel framework for the co-segmentation task. Our network design has two key components: a region correspondence module which is the basic operation for exchanging information between local image regions, and a cycle refinement module, which utilizes ConvLSTMs to progressively update image representations and exchange information in a cycle and iterative manner. Extensive experiments demonstrate that our proposed method significantly outperforms the state-of-the-art methods on four popular benchmark datasets - PASCAL VOC dataset, MSRC dataset, Internet dataset, and iCoseg dataset, by 2.6%, 7.7%, 2.2%, and 2.9%, respectively.

Real-Time Instance Segmentation of Traffic Videos for Embedded Devices

The paper proposes a novel instance segmentation method for traffic videos devised for deployment on real-time embedded devices. A novel neural network architecture is proposed using a multi-resolution feature extraction backbone and improved network designs for the object detection and instance segmentation branches. A novel post-processing method is introduced to ensure a reduced rate of false detection by evaluating the quality of the output masks. An improved network training procedure is proposed based on a novel label assignment algorithm. An ablation study on speed-vs.-performance trade-off further modifies the two branches and replaces the conventional ResNet-based performance-oriented backbone with a lightweight speed-oriented design. The proposed architectural variations achieve real-time performance when deployed on embedded devices. The experimental results demonstrate that the proposed instance segmentation method for traffic videos outperforms the you only look at coefficients algorithm, the state-of-the-art real-time instance segmentation method. The proposed architecture achieves qualitative results with 31.57 average precision on the COCO dataset, while its speed-oriented variations achieve speeds of up to 66.25 frames per second on the Jetson AGX Xavier module.