Lightweight colon polyp segmentation algorithm based on improved DeepLabV3.

To address the problems that the current polyp segmentation model is complicated and the segmentation accuracy needs to be further improved, a lightweight polyp segmentation network model Li-DeepLabV3+ is proposed. Firstly, the optimized MobileNetV2 network is used as the backbone network to reduce the model complexity. Secondly, an improved simple pyramid pooling module is used to replace the original Atrous Spatial Pyramid Pooling structure, which improves the model training efficiency of the model while reducing the model parameters. Finally, to enhance the feature representation, in the feature fusion module, the low-level feature and the high-level feature are fused using the improved Unified Attention Fusion Module, which applies both channel and spatial attention to enrich the fused features, thus obtaining more boundary information. The model was combined with transfer learning for training and validation on the CVC-ClinicDB and Kvasir SEG datasets, and the generalization of the model was verified across the datasets. The experiment results show that the Li-DeepLabV3+ model has superior advantages in segmentation accuracy and segmentation speed, and has certain generalization abilities.

A flexible and visually meaningful multi-image compression, encryption and hiding scheme based on 2D compressive sensing.

Background: Encrypting plain images into noise-like cipher images is a common method in image encryption. However, when noise-like images appear in public networks, they are more likely to attract attention and suffer more cryptanalysis. To solve this problem, researchers propose the concept of visually meaningful image encryption scheme, which encrypts a plain image into a visually meaningful cipher image. Objective: In order to realize the visual security of cipher image and increase information capacity, this paper proposes a flexible visually secure multi-image compression, encryption and hiding scheme based on two-dimensional compressive sensing (2DCS), which can flexibly complete the compression and encryption of multiple plain images without increasing the amount of ciphertext data. Methods: The scheme is divided into encryption process and embedding process. In the encryption process, the plain image is randomly scrambled and non-linear gray value transformed to obtain a pre-encrypted integer matrix, then 2DCS is used to compress the pre-encrypted integer matrix to get the secret image. Repeat this process for multiple plain images to obtain multiple secret images. In the embedding process, integer wavelet transform and bit-plane decomposition are used to embed multiple secret images into the quantized coefficient matrix of host image to get the modified coefficient matrix, and then the inverse integer wavelet transform is used to transform the modified coefficient matrix into spatial space to get the visually meaningful cipher image. Result: The simulation experiment verifies the feasibility and effectiveness of the visually meaningful multi-image encryption scheme, and users can choose to improve the system's encryption capacity or cipher image's visual security according to their own needs.

Dermoscopic Image Classification Method Using an Ensemble of Fine-Tuned Convolutional Neural Networks.

Aiming at the problems of large intra-class differences, small inter-class differences, low contrast, and small and unbalanced datasets in dermoscopic images, this paper proposes a dermoscopic image classification method based on an ensemble of fine-tuned convolutional neural networks. By reconstructing the fully connected layers of the three pretrained models of Xception, ResNet50, and Vgg-16 and then performing transfer learning and fine-tuning the three pretrained models with the ISIC 2016 Challenge official skin dataset, we integrated the outputs of the three base models using a weighted fusion ensemble strategy in order to obtain a final prediction result able to distinguish whether a dermoscopic image indicates malignancy. The experimental results show that the accuracy of the ensemble model is 86.91%, the precision is 85.67%, the recall is 84.03%, and the F1-score is 84.84%, with these four evaluation metrics being better than those of the three basic models and better than some classical methods, proving the effectiveness and feasibility of the proposed method.

Skin Disease Recognition Method Based on Image Color and Texture Features.

Skin diseases have a serious impact on people's life and health. Current research proposes an efficient approach to identify singular type of skin diseases. It is necessary to develop automatic methods in order to increase the accuracy of diagnosis for multitype skin diseases. In this paper, three type skin diseases such as herpes, dermatitis, and psoriasis skin disease could be identified by a new recognition method. Initially, skin images were preprocessed to remove noise and irrelevant background by filtering and transformation. Then the method of grey-level co-occurrence matrix (GLCM) was introduced to segment images of skin disease. The texture and color features of different skin disease images could be obtained accurately. Finally, by using the support vector machine (SVM) classification method, three types of skin diseases were identified. The experimental results demonstrate the effectiveness and feasibility of the proposed method.

Cervical cancer histology image identification method based on texture and lesion area features.

The issue of an automated approach for detecting cervical cancer is proposed to improve the accuracy of recognition. Firstly, the cervical cancer histology source images are needed to use image preprocessing for reducing the impact brought by noise of images as well as the impact on subsequent precise feature extraction brought by irrelevant background. Secondly, the images are grouped into ten vertical images and the information of texture feature is extracted by Grey Level Co-occurrence Matrix (GLCM). GLCM is an effective tool to analyze the features of texture. The textures of different diseases in the source image of Cervical Cancer Histology (such as contrast, correlation, entropy, uniformity and energy, etc.) can all be obtained in this way. Thirdly, the image is segmented by using K-means clustering and Marker-controlled watershed Algorithm. And each vertical image is divided into three layers to calculate the areas of different layers. Based on GLCM and lesion area features, the tissues are investigated with segmentation by using Support Vector Machine (SVM) method. Finally, the results show that it is effective and feasible to recognize cervical cancer by automated approach and verified by experiment.