Real-time vehicle detection, tracking and counting from surveillance cameras is a main part for many applications in smart cities. Usually, this task encounters some problems in practice, like the lack of real-time processing of the videos or the errors in detection and/or tracking. In this paper we propose an approach for real time vehicle counting by using Tiny YOLO for detection and fast motion estimation for tracking. Our application is running in Ubuntu with GPU processing, and the next step is to test it on low-budget devices, as Jetson Nano. Experimental results show that our approach achieves high accuracy at real time speed (33.5 FPS) on real traffic videos.

Liver segmentation from computer tomography scans is a topic of research interest, as the acquisition and inter-patient variability make the automatic segmentation difficult. The current trend is to improve the accuracy and to reduce the computational complexity of the segmentation, as this is essential for the diagnostic and for 3D rendering. We propose a new computationally efficient approach for 3D liver segmentation, based on the 3D Discrete Cosine Transform applied on volume blocks for feature extraction, followed by a support vector machine classification of volume blocks. The segmentation is refined in a post-processing step through a 3D median filtering, 3D morphological operations, and 3D connected components analysis. This new method has been applied on real liver volumes and provided promising results, on the level of the state of the art, with a significant reduction in the data to be processed and in the operations involved as compared to other approaches.

Image filtering is one of the principal tools used in computer vision applications. Real systems store and manipulate high resolution images in compressed forms, therefore the implementation of the entire processing chain directly in the compressed domain became essential. This includes almost always linear filtering operations implemented by convolution. Linear image filtering implementation directly on the JPEG images is challenging for several reasons, including the complexity of transposing the pixel level convolution in the compressed domain, which may increase the processing time, despite avoiding the decompression. In this paper we propose a new computationally efficient solution for JPEG image filtering (as a spatial convolution between the input image and a given kernel) directly in the DCT based compressed domain. We propose that the convolution operation to be applied just on the periodical extensions of the DCT basis images, as an off-line processing, obtaining the filtered DCT basis images, which are used in data decompression. While this doesn't solve the near block boundaries filtering artefacts for large convolution kernels, for most practical cases, it provides good quality results at a very low computational complexity. These kind of implementations can run at real-time rates/ speeds and are suitable for developments of applications on digital cameras/ DSP/FPGA.

Skin moles and lesions can be the first signs of severe skin diseases such as cancer. This paper presents the development of an end-user device capable of capturing images, segmentation and diagnosis of moles by using the ABCD rule, which stands for analyzing moles’ parameters as: asymmetry, border, color, and diameter. These are the main mole characteristics that doctors look at, each of them having a different factor of importance, and depending on these an accurate diagnosis can be given. For the hardware, we developed a small and compact device that can be manipulated easily by anyone without knowledge of medicine, in which we considered a custom-designed 3D enclosure with two white LEDs to control the light. The device has the role of facilitating analysis of the suspicious moles regularly at home, even if only from an indicative and not from a medical point of view. The developed PC software permits the storage of the images in a local database for easy tracking and analysis in time. The image processing developed for the ABCD rule is incorporated into the PC software and tested extensively on the international PH2 database with skin melanoma images to validate our segmentation and criteria evaluation. Using the developed device, we captured mole images for patients, who also took a medical examination by a specialist using the standard dermatoscope. Therefore, we obtained our own database containing 26 images for which we have also the specialists’ diagnosis. The performance evaluation measures obtained using our device are—Accuracy: 0.92, Precision: 1.0, Recall: 0.92, F1-score: 0.96.