Lyme rashes disease classification using deep feature fusion technique

Automatic classification of Lyme disease rashes on the skin helps clinicians and dermatologists' probe and investigate Lyme skin rashes effectively. This paper proposes a new in-depth features fusion system to classify Lyme disease rashes. The proposed method consists of two main steps. First, three different deep learning models, Densenet201, InceptionV3, and Exception, were trained independently to extract the deep features from the erythema migrans (EM) images. Second, a deep feature fusion mechanism (meta classifier) is developed to integrate the deep features before the final classification output layer. The meta classifier is a basic deep convolutional neural network trained on original images and features extracted from base level three deep learning models. In the feature fusion mechanism, the last three layers of base models are dropped out and connected to the meta classifier. The proposed deep feature fusion method significantly improved the classification process, where the classification accuracy was 98.97%, which is particularly impressive than the other state-of-the-art models.

The impact of fusion imaging technique on middle ear cholesteatoma surgery: a prospective comparative study

BACKGROUND

The fusion of computed tomography images with non-echo planner diffusion-weighted magnetic resonance images may overcome the limitations of each individual modality.

OBJECTIVES

This study aimed to assess the ability of the 'fusion' technique to predict the location of middle ear cholesteatoma by evaluating Its impact on preoperative surgical planning and postoperative results.

METHODOLOGY

Eighty-three adults with cholesteatoma underwent preoperative CT scans and non-EPI-DW-MRI with or without the 'fusion' technique. We evaluated cholesteatoma localization in both groups, selecting the most appropriate surgical technique, and correlating it with the intraoperative findings. Both groups were compared in terms of residual/recurrent cholesteatoma at one, six and twelve months after surgery.

RESULTS

The 'fusion' technique's sensitivity, specificity, and accuracy in predicting the location of cholesteatoma were 97.5. 97.4, and 97.5%, respectively, versus 97.59, 57.69, and 73.21% of CT alone. There was a total operators agreement after the evaluation of the 'fusion' images with the adopted surgical technique. A statistically significant decrease in residual and recurrence cholesteatoma rates was found in group A.

CONCLUSIONS

The 'fusion' technique provides the surgeon with the precise cholesteatoma location, guiding him in making the correct surgical decision, contributing to the decrease in postoperative residual and recurrence rates.

Paravalvular leak closure with real time transesophageal echocardiography and fluoroscopy fusion

Transcatheter technology has been increasingly used for paravalvular leak closure. We report the use of “Fusion Technology” (EchoNaviagator, Phillips, Tustin, CA) that combines real-time 2 and 3 dimensional trans-esophageal echocardiography with fluoroscopy imaging to facilitate paravalvular leak closure. This could help to identify the exact site, size, depth and shape of the paravalvular leak for proper positioning of the occluder device, which may result in saving time and effort.

Color Image Generation from Range and Reflection Data of LiDAR

Recently, it has been reported that a camera-captured-like color image can be generated from the reflection data of 3D light detection and ranging (LiDAR). In this paper, we present that the color image can also be generated from the range data of LiDAR. We propose deep learning networks that generate color images by fusing reflection and range data from LiDAR point clouds. In the proposed networks, the two datasets are fused in three ways-early, mid, and last fusion techniques. The baseline network is the encoder-decoder structured fully convolution network (ED-FCN). The image generation performances were evaluated according to source types, including reflection data-only, range data-only, and fusion of the two datasets. The well-known KITTI evaluation data were used for training and verification. The simulation results showed that the proposed last fusion method yields improvements of 0.53 dB, 0.49 dB, and 0.02 in gray-scale peak signal-to-noise ratio (PSNR), color-scale PSNR, and structural similarity index measure (SSIM), respectively, over the conventional reflection-based ED-FCN. Besides, the last fusion method can be applied to real-time applications with an average processing time of 13.56 ms per frame. The methodology presented in this paper would be a powerful tool for generating data from two or more heterogeneous sources.