Computer-aided early diagnosis of Alzheimers Disease (AD) and its prodromal form, Mild Cognitive Impairment (MCI), has been the subject extensive research in recent years. Some studies have shown promising results AD MCI determination using structural functional Magnetic Resonance Imaging (sMRI, fMRI), Positron Emission Tomography (PET) Diffusion Tensor (DTI) modalities. Furthermore, fusion imaging modalities a supervised machine learning framework direction research. In this paper we first...

The pervasion of 3-D technologies over the years gives rise to increasing demands accurate and efficient stereoscopic image quality assessment (SIQA) methods, designed automatically supervise optimize video processing systems. Though 2-D IQA has attracted considerable attention, its counterpart is yet be well explored. In this paper, a no-reference SIQA method using convolution neural network (CNN) for feature extraction proposed. proposed method, CNN model trained from scratch classify...

The most recent 3D object detectors for point clouds rely on the coarse voxel-based representation rather than accurate point-based due to a higher box recall in Region Proposal Network (RPN). However, detection accuracy is severely restricted by information loss of pose details voxels. Different from considering cloud as voxel or only, we propose point-to-voxel feature learning approach voxelize with both point-wise semantic and local spatial features, which maintains voxel-wise features...

Most existing point cloud based 3D object detectors focus on the tasks of classification and box regression. However, another bottleneck in this area is achieving an accurate detection confidence for Non-Maximum Suppression (NMS) post-processing. In paper, we add a IoU prediction branch to regular regression branches. The predicted used as NMS. order obtain more prediction, propose IoU-Net with sensitive feature learning alignment operation. To perspective-invariant head, Attentive Corner...

The two-dimensional multiwindow S-method for radar imaging applications is proposed. It represents a combined technique that uses the standard and multiple windows approach based on Hermite functions. proposed method provides significant improvement of image concentration in comparison with S-method. Also, it does not require an additional post-processing algorithm. efficiency demonstrated through various examples.

The perceptual quality of stereoscopic images plays an essential role in the human perception visual information. However, most available image assessment (SIQA) methods evaluate 3D experience using hand-crafted features or shallow architectures, which cannot model properties stereo well. In this paper, we use convolutional neural networks (CNNs) to learn deeper local quality-aware structures for images. With different inputs, two CNN models are designed no-reference SIQA tasks. one-column...

A method for temporal analysis and reconstruction of video sequences based on the time-frequency Hermite projection is proposed. The S-method-based distribution used to characterize stationarity within sequence. Namely, a sequence DCT coefficients along time axes create frequency-modulated signal. nonstationary done using expansion coefficients. Here, small number can be used, which may provide significant savings some video-based applications. results are illustrated with examples.

In this work we develop a post-processing algorithm which enhances the results of existing image deblurring methods. It performs additional edge sharpening using grid warping. The idea proposed is to transform neighborhood so that neighboring pixels move closer edge, and then resample from warped original uniform grid. technique preserves textures while making edges sharper. effectiveness method shown for basic methods on LIVE database images with added blur noise.

In this work we propose a post-processing method for BM3D algorithm that has become state-of-the-art image denoising and deblurring algorithm. Although produces results with high objective metrics values, it also adds noticeable high-frequency artifacts. We suppress these artifacts using second order Total Generalized Variation (TG V) TGV is an extension of but does not tend to make images piecewise constant. suggest efficient numerical scheme minimization. validate the proposed idea, tests...

Suppression of ringing effect is a challenging problem. It mainly caused by absence effective methods artifact detection. In this paper we introduce estimation method based on scale-space analysis. The shows good results for low-pass filtered test images and in adaptive image deringing.

Motivated by the success of convolutional neural networks (CNNs) in image-related applications, this paper, we design an effective method for no-reference 3D image quality assessment (3D IQA) through CNN-based feature extraction and consolidation strategy. In first most vital stage, quality-aware features, which reflect inherent images, are extracted a fine-tuned CNN model exploiting concept transfer learning. This fine-tuning strategy solves large-scale training data dependence existing...

This paper presents a new adaptive post-processing algorithm for ringing artifact reduction after image interpolation (up sampling). The is based on the concept of total variation (TV) control. It uses known TV blocks low-resolution image. Conditional gradient, subgradient and projection methods this are considered analyzed. A test set 181300 overlapping 11times11 real images was used local optimization analysis. Local conditional gradient method shows best objective subjective results.