Variable weight neural networks and their applications on material surface and epilepsy seizure phase classifications

HWA-SegNet: Multi-channel skin lesion image segmentation network with hierarchical analysis and weight adjustment

Convolutional neural networks (CNNs) show excellent performance in accurate medical image segmentation. However, the characteristics of sample with small size and insufficient feature expression, irregular shape of the segmented target and inaccurate judgment of edge texture have always been problems to be faced in the field of skin lesion image segmentation. Therefore, in order to solve these problems, discrete Fourier transform (DFT) is introduced to enrich the input data and a CNN architecture (HWA-SegNet) is proposed in this paper. Firstly, DFT is improved to analyze the features of the skin lesions image, and multi-channel data is extended for each image. Secondly, a hierarchical dilated analysis module is constructed to understand the semantic features under multi-channel. Finally, the pre-prediction results are fine-tuned using a weight adjustment structure with fully connected layers to obtain higher accuracy prediction results. Then, 520 skin lesion images are tested on the ISIC 2018 dataset. Extensive experimental results show that our HWA-SegNet improve the average segmentation Dice Similarity Coefficient from 88.30% to 91.88%, Sensitivity from 89.29% to 92.99%, and Jaccard similarity index from 81.15% to 85.90% compared with U-Net. Compared with the State-of-the-Art method, the Jaccard similarity index and Specificity are close, but the Dice Similarity Coefficient is higher. The experimental data show that the data augmentation strategy based on improved DFT and HWA-SegNet are effective for skin lesion image segmentation.

Compensating Uncertainties in Force Sensing for Robotic-Assisted Palpation

Force sensing in robotic-assisted minimally invasive surgery (RMIS) is crucial for performing dedicated surgical procedures, such as bilateral teleoperation and palpation. Due to the bio-compatibility and sterilization requirements, a specially designed surgical tool/shaft is normally attached to the sensor while contacting the organ targets. Through this design, the measured force from the sensor usually contains uncertainties, such as noise, inertial force etc., and thus cannot reflect the actual interaction force with the tissue environment. Motivated to provide the authentic contact force between a robotic tool and soft tissue, we proposed a data-driven force compensation scheme without intricate modeling to reduce the effects of force measurement uncertainties. In this paper, a neural-network-based approach is utilized to automatically model the inertial force subject to noise during the robotic palpation procedure, then the exact contact force can be obtained through the force compensation method which cancels the noise and inertial force. Following this approach, the genuine interaction force during the palpation task can be achieved furthermore to improve the appraisal of the tumor surrounded by the soft tissue. Experiments are conducted with robotic-assisted palpation tasks on a silicone-based soft tissue phantom and the results verify the effectiveness of the suggested method.