Traditional radiomics models mainly rely on explicitly-designed handcrafted features from medical images. This paper aimed to investigate if deep features extracted via transfer learning can generate radiomics signatures for prediction of overall survival (OS) in patients with Glioblastoma Multiforme (GBM). This study comprised a discovery data set of 75 patients and an independent validation data set of 37 patients. A total of 1403 handcrafted features and 98304 deep features were extracted from preoperative multi-modality MR images. After feature selection, a six-deep-feature signature was constructed by using the least absolute shrinkage and selection operator (LASSO) Cox regression model. A radiomics nomogram was further presented by combining the signature and clinical risk factors such as age and Karnofsky Performance Score. Compared with traditional risk factors, the proposed signature achieved better performance for prediction of OS (C-index = 0.710, 95% CI: 0.588, 0.932) and significant stratification of patients into prognostically distinct groups (P < 0.001, HR = 5.128, 95% CI: 2.029, 12.960). The combined model achieved improved predictive performance (C-index = 0.739). Our study demonstrates that transfer learning-based deep features are able to generate prognostic imaging signature for OS prediction and patient stratification for GBM, indicating the potential of deep imaging feature-based biomarker in preoperative care of GBM patients.

Performance of blind image quality assessment (BIQA) models has been significantly boosted by end-to-end optimization of feature engineering and quality regression. Nevertheless, due to the distributional shift between images simulated in the laboratory and captured in the wild, models trained on databases with synthetic distortions remain particularly weak at handling realistic distortions (and vice versa). To confront the cross-distortion-scenario challenge, we develop a unified BIQA model and an approach of training it for both synthetic and realistic distortions. We first sample pairs of images from individual IQA databases, and compute a probability that the first image of each pair is of higher quality. We then employ the fidelity loss to optimize a deep neural network for BIQA over a large number of such image pairs. We also explicitly enforce a hinge constraint to regularize uncertainty estimation during optimization. Extensive experiments on six IQA databases show the promise of the learned method in blindly assessing image quality in the laboratory and wild. In addition, we demonstrate the universality of the proposed training strategy by using it to improve existing BIQA models.

The topics of visual and audio quality assessment (QA) have been widely researched for decades, yet nearly all of this prior work has focused only on single-mode visual or audio signals. However, visual signals rarely are presented without accompanying audio, including heavy-bandwidth video streaming applications. Moreover, the distortions that may separately (or conjointly) afflict the visual and audio signals collectively shape user-perceived quality of experience (QoE). This motivated us to conduct a subjective study of audio and video (A/V) quality, which we then used to compare and develop A/V quality measurement models and algorithms. The new LIVE-SJTU Audio and Video Quality Assessment (A/V-QA) Database includes 336 A/V sequences that were generated from 14 original source contents by applying 24 different A/V distortion combinations on them. We then conducted a subjective A/V quality perception study on the database towards attaining a better understanding of how humans perceive the overall combined quality of A/V signals. We also designed four different families of objective A/V quality prediction models, using a multimodal fusion strategy. The different types of A/V quality models differ in both the unimodal audio and video quality prediction models comprising the direct signal measurements and in the way that the two perceptual signal modes are combined. The objective models are built using both existing state-of-the-art audio and video quality prediction models and some new prediction models, as well as quality-predictive features delivered by a deep neural network. The methods of fusing audio and video quality predictions that are considered include simple product combinations as well as learned mappings. Using the new subjective A/V database as a tool, we validated and tested all of the objective A/V quality prediction models. We will make the database publicly available to facilitate further research.

Audio information has been bypassed by most of current visual attention prediction studies. However, sound could have influence on visual attention and such influence has been widely investigated and proofed by many psychological studies. In this paper, we propose a novel multi-modal saliency (MMS) model for videos containing scenes with high audio-visual correspondence. In such scenes, humans tend to be attracted by the sound sources and it is also possible to localize the sound sources via cross-modal analysis. Specifically, we first detect the spatial and temporal saliency maps from the visual modality by using a novel free energy principle. Then we propose to detect the audio saliency map from both audio and visual modalities by localizing the moving-sounding objects using cross-modal kernel canonical correlation analysis, which is first of its kind in the literature. Finally we propose a new two-stage adaptive audiovisual saliency fusion method to integrate the spatial, temporal and audio saliency maps to our audio-visual saliency map. The proposed MMS model has captured the influence of audio, which is not considered in the latest deep learning based saliency models. To take advantages of both deep saliency modeling and audio-visual saliency modeling, we propose to combine deep saliency models and the MMS model via a later fusion, and we find that an average of 5% performance gain is obtained. Experimental results on audio-visual attention databases show that the introduced models incorporating audio cues have significant superiority over state-of-the-art image and video saliency models which utilize a single visual modality.

Vision-based intelligent systems like automatic driving or driving assistance can be improved by enhancing the visibility of the scenes captured in bad weather conditions. In particular, many image dehazing algorithms (DHAs) have been proposed to facilitate such applications in hazy weather. Contrary to the substantial progress of DHA developing, the quality evaluation of DHAs falls behind. Generally, DHAs can be evaluated qualitatively by human subjects or quantitatively by objective quality measures. Compared with the subjective evaluation which is time consuming and difficult to apply, objective measures with quantitative results are more needed in practical systems. But in the literature, very few measures are widely utilized, and even less measures correlate well with the overall dehazing quality (DHQ). In this paper, we study the DHQ evaluation using real hazy images systematically. We first construct a DHQ database, which is the largest of its kind so far and includes 1750 dehazed images generated from 250 real hazy images of various haze densities using seven representative DHAs. A subjective quality evaluation study is subsequently conducted on the DHQ database. Then, we propose an objective DHQ index (DHQI) by extracting and fusing three groups of features, including: 1) haze-removing features; 2) structure-preserving features; and 3) over-enhancement features, which have captured the most key aspects of dehazing. DHQI can be utilized to evaluate DHAs or optimize practical dehazing systems. Validations on the constructed DHQ database and three other databases with synthetic haze have verified the effectiveness of DHQI. Finally, we give an overview of the current DHA quality evaluation strategies, discuss their merits and demerits, and give some suggestions on systematic DHA quality evaluation. The DHQ database and the code of DHQI will be released to facilitate further research.