Software & Code
Orasis Image Enhancement: Bridging the gap between what you see and what the camera captures
Orasis is a biologically-inspired, image enhancement software, which employs the characteristics of the ganglion cells of the Human Visual System. Many times the image captured by a camera and the image in our eyes are dramatically different. Especially when there are shadows or highlights in the same scene. In these cases our eyes can distinguish many more details in the shadows or highlights, while the image captured by the camera suffers from loss of visual information in these regions.
Orasis attempts to bridge the gap between "what you see" and "what the camera outputs". It enhances the shadow or the highlight regions of an image, while keeping intact all the correctly exposed ones. The final result is a lot closer to the human perception of the scene, than the original captured image, revealing visual information that otherwise wouldn't be available to the human observer. Additionally to the above, Orasis can correct low local contrast, colors and noise.
Orasis for iOS
Orasis is available for iPhone & iPad.
Official Orasis iOS website: https://www.orasisapp.com/
Orasis in the AppStore:
OrasisHD (paid, full functionality): https://apps.apple.com/us/app/orasishd/id454408758
Orasis Lite (free, test functionality): https://apps.apple.com/us/app/orasis/id1050740833
Orasis for Windows
The following video demonstrates the use of the Orasis desktop application. It is developed in Visual Studio C# and is available for Windows platforms.
Free PYTHON code for image enhancement
Free MATLAB code for HDR Multi-Exposure Image Fusion
Free PYTHON / MATLAB code for dataset shaping and balancing
Scripts for replicating the results
A generic function that can be used for shaping (or balancing) datasets
If you use this code in your research, please cite the following papers:
Vonikakis, V., Subramanian, R., Arnfred, J., & Winkler, S. A Probabilistic Approach to People-Centric Photo Selection and Sequencing. IEEE Transactions in Multimedia, 11(19), pp.2609-2624, 2017.
V. Vonikakis, R. Subramanian, S. Winkler. (2016). Shaping Datasets: Optimal Data Selection for Specific Target Distributions. Proc. ICIP2016, Phoenix, USA, Sept. 25-28.
Free PYTHON code for facial expression analysis (with DLIB)
Free Python code repo for Dimensional Facial Expression Analysis (estimation of Arousal, Valence, Intensity) from facial landmarks extracted with DLIB. https://github.com/bbonik/facial-expression-analysis
Free PYTHON code for 2D facial landmark frontalization (extracted by DLIB)
In this repository, you can find a Python function that can easily frontalize 2D facial landmarks extracted by the DLIB library. There are examples for applying the function in static images and camera video stream.
Free PYTHON code for robust document binarization
Python function for document binarization, inspired by the OFF center-surround cells of the Human Visual System. The function is particularly good for strong intensity variations, like stains or shadows, and thus good for binarizing text that is captured outdoors, e.g. using a mobile phone, or binarizing license plates.
You can find this function, along with examples and comparisons with other methods, in this Github repository.
Library for Machine Learning model performance reporting
In this repository, you can find Python functions for generating nice looking performance reports for your ML models.
It includes confusion matrices, ROC and PR curves, TPR/FPR/TNR/FNR and threshold analysis for binary classifiers.
Library for estimating glare on a formed retinal image
In this repository, you can find Python and Matlab functions for estimating the impact of glare on the formed retinal image, derived from derived from calibrated measurements of scene radiances, or luminances.
Automatic Slideshow Creation / Image Appeal Measure / Image Selection Probabilities
This is the supplementary material for the paper: "A Probabilistic Approach to People-Centric Image Selection and Sequencing".
The image selection probabilities for 13 image/face attributes, learnt through a large-scale crowdsourcing study.
A MATLAB implementation of the Mixed-Integer Linear Programming (MILP) technique for dataset shaping (or balancing), including minimization of the cross-dimensional correlations.
A MATLAB implementation of the Integer Linear Programming (ILP) technique for automatic appealing slideshow creation, based on the learnt selection probabilities from the crowd.
The code can be downloaded from here