Forgery Detection Using Inconsistent Reflection Vanishing Point

The rapid development of image processing techniques has made it extremely easy to alter the content of images or create newimages.So photographs,which appear in magazines,social media,and political attacks,can no longer be trusted.A novel and effective technique is proposed in this paper to expose image forgery using inconsistent reflection vanishing point(RVP).More specifically,the definition of error distance is given,sin^2()-based function is proposed to normalize error distance,and a reasonable threshold value is set to detect image forgery.The experimental data and results are presented to demonstrate the accuracy and effectiveness of the technique.

A Study on Fast and Robust Vanishing Point Detection System Using Fast M-Estimation Method and Regional Division for In-vehicle Camera

The vanishing point detection technology helps automatic driving. In this paper, the straight lines on the road associated with the vanishing point are extracted efficiently by using the regional division and angle limitation. And, the vanishing point is detected robustly by using the fast M-estimation method. Proposed method could detect straight-line features associated with vanishing point detection efficient on the road. And the vanishing point was detected exactly by the effect of the fast M-estimation method when the straight-line features not associated with vanishing point detection were detected. The processing time of the proposed method was faster than the camera flame rate （30 fps）. Thus, the proposed method is capable of real-time processing.

Vanishing Point

Reading culture in many African societies is on a downward spiral READING culture varies in a number of ways across African societies,but what is generally perceived about Africans is that they are not noted for frequently reading books,magazines,newspapers or any literary materials-even in their leisure time. This is summed up in a popular adage,which says that whenever you want to hide something from an African,you should just write it in a book. But is this perception really true that Africans are people who lack a good reading culture?

VCANet:Vanishing-Point-Guided Context-Aware Network for Small Road Object Detection

Advanced deep learning technology has made great progress in generic object detection of autonomous driving,yet it is still challenging to detect small road hazards in a long distance owing to lack of large-scale small-object datasets and dedicated methods.This work addresses the challenge from two aspects.Firstly,a self-collected long-distance road object dataset(TJ-LDRO)is introduced,which consists of 109,337 images and is the largest dataset so far for the small road object detection research.Secondly,a vanishing-point-guided context-aware network(VCANet)is proposed,which utilizes the vanishing point prediction block and the context-aware center detection block to obtain semantic information.The multi-scale feature fusion pipeline and the upsampling block in VCANet are introduced to enhance the region of interest(ROI)feature.The experimental results with TJ-LDRO dataset show that the proposed method achieves better performance than the representative generic object detection methods.This work fills a critical capability gap in small road hazards detection for high-speed autonomous vehicles.

Abnormal driving behavior identification based on direction and position offsets

Abnormal driving behavior identification( ADBI) has become a research hotspot because of its significance in driver assistance systems. However,current methods still have some limitations in terms of accuracy and reliability under severe traffic scenes. This paper proposes a new ADBI method based on direction and position offsets,where a two-factor identification strategy is proposed to improve the accuracy and reliability of ADBI. Self-adaptive edge detection based on Sobel operator is used to extract edge information of lanes. In order to enhance the efficiency and reliability of lane detection,an improved lane detection algorithm is proposed,where a Hough transform based on local search scope is employed to quickly detect the lane,and a validation scheme based on priori information is proposed to further verify the detected lane. Experimental results under various complex road conditions demonstrate the validity of the proposed ADBI.

Road Geometry Recovery Using On-board Camera in Various Conditions

A 3D road geometry reconstruction system using on-board camera for vehicle driver assistance is developed. Given calibration of the camera, 3D structure of the road in front of the vehicle is built via multiple vanishing points estimation and local parallelism of the road. The road will be modeled as a piecewise linear approximation of space curve. Vanishing points are estimated through line grouping and histogram method in the image. The system is not restricted to the roads which lie on 2D plane, but also is capable of recovering a full 3D geometry. The ambiguity of distance （scale） is resolved by usage of parallel line segnents an the road image.

Exposing photo manipulation with inconsistent perspective geometry

Manipulated digital image is got interesting in recent years. Digital images can be manipulated more easily with the aid of powerful image editing software. Forensic techniques for authenticating the integrity of digital images and exposing forgeries are urgently needed. A geometric-based forensic technique which exploits the principle of vanishing points is proposed. By means of edge detection and straight lines extraction, intersection points of the projected parallel lines are computed. The normalized mean value （NMV） and normalized standard deviation （NSD） of the distances between the intersection points are used as evidence for image forensics. The proposed method employs basic rules of linear perspective projection, and makes minimal assumption. The only requirement is that the parallel lines are contained in the image. Unlike other forensic techniques which are based on low-level statistics, this method is less sensitive to image operations that do not alter image content, such as image resampling, color manipulation, and lossy compression. This method is demonstrated with images from York Urban database. It shows that the proposed method has a definite advantage at separating authentic and forged images.