Photograph manipulation | EPFL Graph Search

Photograph manipulation involves the of a photograph. Some photograph manipulations are considered to be skillful artwork, while others are considered to be unethical practices, especially when used to deceive. Photographs may be manipulated for political propaganda, to improve the appearance of a subject, for entertainment, or as humor. Depending on the application and intent, some photograph manipulations are considered an art form because they involve creation of unique images and in some instances, signature expressions of art by photographic artists. For example, Ansel Adams used darkroom exposure techniques, burning (darkening) and dodging (lightening) a photograph. Other techniques include retouching using ink or paint, airbrushing, double exposure, piecing photos or negatives together in the darkroom, and scratching instant films. Software tools applied to digital images range from professional applications to basic imaging software for casual users. Photoshopping is a verb

Privacy-Friendly Photo Sharing and Relevant Applications Beyond

Lin Yuan

Popularization of online photo sharing brings people great convenience, but has also raised concerns for privacy. Researchers proposed various approaches to enable image privacy, most of which focus on encrypting or distorting image visual content. In this thesis, we investigate novel solutions to protect image privacy with a particular emphasis on online photo sharing. To this end, we propose not only algorithms to protect visual privacy in image content but also design of architectures for privacy-preserving photo sharing. Beyond privacy, we also explore additional impacts and potentials of employing daily images in other three relevant applications. First, we propose and study two image encoding algorithms to protect visual content in image, within a Secure JPEG framework. The first method scrambles a JPEG image by randomly changing the signs of its DCT coefficients based on a secret key. The second method, named JPEG Transmorphing, allows one to protect arbitrary image regions with any obfuscation, while secretly preserving the original image regions in application segments of the obfuscated JPEG image. Performance evaluations reveal a good degree of storage overhead and privacy protection capability for both methods, and particularly a good level of pleasantness for JPEG Transmorphing, if proper manipulations are applied. Second, we investigate the design of two architectures for privacy-preserving photo sharing. The first architecture, named ProShare, is built on a public key infrastructure (PKI) integrated with a ciphertext-policy attribute-based encryption (CP-ABE), to enable the secure and efficient access to user-posted photos protected by Secure JPEG. The second architecture is named ProShare S, in which a photo sharing service provider helps users make photo sharing decisions automatically based on their past decisions using machine learning. The photo sharing service analyzes not only the content of a user's photo, but also context information about the image capture and a prospective requester, and finally makes decision whether or not to share a particular photo to the requester, and if yes, at which granularity. A user study along with extensive evaluations were performed to validate the proposed architecture. In the end, we research into three relevant topics in regard to daily photos captured or shared by people, but beyond their privacy implications. In the first study, inspired by JPEG Transmorphing, we propose an animated JPEG file format, named aJPEG. aJPEG preserves its animation frames as application markers in a JPEG image and provides smaller file size and better image quality than conventional GIF. In the second study, we attempt to understand the impact of popular image manipulations applied in online photo sharing on evoked emotions of observers. The study reveals that image manipulations indeed influence people's emotion, but such impact also depends on the image content. In the last study, we employ a deep convolutional neural network (CNN), the GoogLeNet model, to perform automatic food image detection and categorization. The promising results obtained provide meaningful insights in design of automatic dietary assessment system based on multimedia techniques, e.g. image analysis.

EPFL2017

Foot-Interfaces Control and Human-Robot Interaction in Four-Handed Manipulation

Jacob Hernandez Sanchez

Humans have a great capacity to perform complex manipulations. However, there are many tasks for which two hands are not enough, such as in surgery. For such tasks, a dyad of people is required. However, when not used to work together, a dyad collaboration can be intermittent, slow and prone to misunderstandings if relied solely on verbal cues. The current PhD thesis report concerns the design and control of foot interfaces to enable solo four-handed manipulations. In the first part, we present the platform design in five DoF, with position mapping and haptic force feedback. We tackle the known issue of fatigue when using position mapping, by partially compensating for the dynamics of the leg. Finally, we perform a series of feasibility validations on bipedal teleoperation in 3 and 4 and 5 DoF. In the second part, a behavioural study investigates a symmetric bipedal holding and transport of an object while working on it with the hands. This study sheds light to the human capacity to use the redundancy given by augmentation with the feet, to assist the task of the hands. We investigate the effect of shared-control for lower limb teleoperation. Two modalities are used: 1. synergistic control of two robotic arms with one foot, 2. force assistance. We evaluate these strategies in terms of objective and subjective metrics, such as physical and cognitive workload and fluency.Aiming at enabling individualized foot dexterous control, in the third part of the thesis, we addressed the problem of bio-mechanical coupling of foot rotations. This is done by detecting intention to grasp, and selectively easing this action while haptically constraining the foot gestures for tool alignment. A user study was conducted to evaluate our proposed approach in a surgical laparoscopic gripper.In the fourth and final part, we focus on investigating more demanding tasks for the feet, as well as the coordinated control of four hands, during hybrid-robotic laparoscopic surgery. One foot controls a camera, while the other one performs a grasping task. Noteworthy is that all four-arms can be controlled simultaneously. For easing the solo surgical task, we design haptic assistance, for autonomous grasping operated by the foot. Finally, we perform a systematic behavioural study to understand the effect of our haptic-shared-control, along with an assistance modality for camera tool-tracking. We cover uni-pedal, bipedal and four-handed task scenarios. Our main conclusion is that foot gestures can be conditioned with haptic feedback, to effectively control multiple degrees of freedom of a supernumerary robotic arm. Thereby enabling four-arm simultaneous manipulations. This is done specifically by leveraging 1. a foot kinesthetic haptic platform that allows for the use of two feet, 2. leg compensation, and virtual assistive impedance, which are useful to increase comfort and alleviate fatigue. 3. Hybridization of control mappings for foot gestures, with limited range of motion and oblique axes of rotation, 4. Haptic shared-control strategies for prehensile force assistance, coordination assistance, and DoF-individuation assistance. We found that robotic assistance is highly appreciated, and has an important role in simplifying four-handed tasks. Indeed, these are instrumental considerations to enable a more fluent, less demanding, and more performant interaction, in tasks where the feet are supplementing the task of hands.

EPFL2022

Evaluation and Prediction of Evoked Emotions Induced by Image Manipulations

Touradj Ebrahimi, Lin Yuan

Various image editing tools make our pictures more attractive, and at the same time, evoke different emotional responses. With powerful and easy-to-use imaging applications, capturing, editing and then sharing pictures have become daily life for many. This paper investigates the influence of several image manipulations on evoked emotions for different types of images. To do so, various types of images clustered in different categories, were collected from Instagram and subjective evaluations were conducted via crowdsourcing to gather the emotional responses on different manipulations as perceived by subjects. Evaluation results show that certain image manipulations can induce different evoked emotions on transformed pictures when compared to the original ones. However, such changes in image emotions due to manipulation are highly content dependent. Then, we conducted a machine learning based experiment, in attempt to predict the emotions of a manipulated image given its original version and the desired manipulation method. Experimental results present a promising performance of such a prediction model, which could pave the road to automatic selection or recommendation of image editing tools that can efficiently transform or emphasize desired emotions in pictures.