Artificial Intelligence Products!
Welcome to Our Artificial Intelligence (AI) Products!
Automatic Kappa angle calculation for patients with nystagmu
The identification and compensation of angle kappa is crucial for the centration of the operative zone in corneal refractive surgeries. However, it still relies on the surgeons’ subjective estimation rather than the precise measurement to compensate for the angle kappa. This may expose potential issues in prior studies of the centration technique as well as affect actual surgical outcomes. Therefore, we have developed a method to identify angle kappa with image recognition with computer vision techniques.
Here, we propose a joint elliptical arc, circular arc, and line segment detector based on the a contrario statistical approach. Our method is an extension of the ELSDc method, recently proposed for line segment and elliptical arc detection. The main contribution is a more general geometrical model, which allows the joint evaluation of the best combination of elliptical arcs, circular arcs, and line segments that corresponds to a given contour. Different interpretations in terms of these elements are tried for the whole contour, instead of locally as it is done in ELSDc. In addition, several minor improvements were performed to the heuristic algorithm used to propose candidates. The performance of the proposed method is compared to the original one on synthetic and real images.
Implementation insights of a pupil- and glint-detection algorithm for high-speed web-based eye tracking. Some of the main metrics are: Movement measures (Amplitude, Saccade peak velocity, Phasic pupil diameter, Tonic pupil diameter), Position measures, Numerosity measures (Fixation count, Regressive fixation count, Blink rate or inter-blink interval) and Latency measures (Fixation duration, Blink amplitude and blink duration).
3D human body scanner
Human body measurement consists of shape and size is a key issue in the various applications that traditionally done by hand. Recently, contactless 3D body surface scanners, based on machine vision technology, are transforming the ability to accurately measure a person’s body size, shape, and skin-surface area.
Weed control equipment
To complete weed control tasks, a special weed mower equipment is produced that can be used and operated via tractors. To be able to utilized in inter-row or intra-row weed removal, machine vision technology has been applied to the field to separate crops from weeds. The digital camera that is located in front of this machine captures images of growth areas under the tractor in real time and uses HSV, adaptive threshold and machine learning methods to distinguish crops from weed areas. In addition, a mechanical method in the form of vibrating teeth weed-fix cultivator is then used to destroy the weeds.
Depth image-based rendering techniques for multiview applications have been recently introduced for efficient view generation at arbitrary camera positions. The rate control in an encoder has thus to consider both texture and depth data. However, due to different structures of depth and texture data and their different roles on the rendered views, the allocation of the available bit budget between them requires a careful analysis. Information loss due to texture coding affects the value of pixels in synthesized views, while errors in depth information lead to a shift in objects or to unexpected patterns at their boundaries. We address the problem of efficient bit allocation between texture and depth data of multiview sequences. We adopt a rate-distortion framework based on a simplified model of depth and texture images, which preserves the main features of depth and texture images. Unlike most recent solutions, our method avoids rendering at encoding time for distortion estimation so that the encoding complexity stays low.
Automatic visual inspection machine
Due to the increasing complexity of assembled printed circuit boards and the demand by end users for fully inspected boards there is now an increasing requirement for automatic optical inspection on assembled boards. Visual inspection of PCB production lines consists of two parts:
1- Solder Paste Inspection (SPI) which is a key inspection equipment to control the solder paste printing quality in SMT. SPI can inspect the solder paste, such as volume, area, height, XY offset, shape, bridge, etc. on PCB.
2- Post-production automated optical inspection (AOI) which is an automated visual inspection of printed circuit board (PCB) (or LCD, transistor) manufacture where a camera autonomously scans the device under test for both catastrophic failure (e.g. missing component) and quality defects (e.g. fillet size or shape or component skew) comparing to an original previously scanned board.
Frontalization and alignment in automatic face detection for people with visual impairments
A system for facilitating a blind person to interact with other people in a way similar to the one with normal vision. The users is the one of a blind person who needs to meet one of his/her acquaintances in a public place, and is not willing to wait for the acquaintance to engage interaction e.g. by speaking: the users prefer to autonomously recognize the person they are meeting, in order to be able to behave consequently.
Optical sorting machine
The automatic optical sorting machine sorts and grades fruits and vegetables according to their shape, color and size. It can sort a wide variety of fruits and vegetables, including: potato, tomato, apple, orange, cherry, etc. quickly and effectively. In addition, the sorting degree based on visual defects is completely adjustable with respect to the shape, color and size of the intended fruit or vegetable. A future development of the machine benefits from internal defect detection of insect infestation using non-destructive IR imaging.
The inefficiency of separable wavelets in representing smooth edges has led to a great interest in the study of new 2-D transformations. The most popular criterion for analyzing these transformations is approximation power. Transformations with near-optimal approximation power are useful in many applications such as denoising and enhancement. However, they are not necessarily good for compression. Therefore, most nearly optimal transformations such as curvelets and contourlets have not yet found any application in image compression. One of the most promising schemes for image compression is the elegant idea of directional wavelets (DIWs). While these algorithms outperform the state-of-the-art image coders in practice, our theoretical understanding of them is very limited. Here, we adopt the notion of rate-distortion and calculate the performance of the DIW on a class of edge-like images. Our theoretical analysis shows that if the edges are not “sharp,” the DIW will compress them more efficiently than the separable wavelets. It also demonstrates the inefficiency of the quadtree partitioning that is often used with the DIW. To solve this issue, we propose a new partitioning scheme called megaquad partitioning.
Straight subjective contour detector
Subjective contours or illusory contours are an important aspect of human perception. Along subjective contours, image contrast is very weak or completely missing, so that no local edge detector can recover them. Their perception is induced by the presence of small pieces of edges and of tips of other long edges incident on the contour. Indeed, in real-world images, edge information of foreground objects is often partly missing due to poor contrast of the object with respect to its background. Nevertheless, the object contour is still perceived by the presence of object or background details that end up abruptly along the contour. Here, we handle the detection of straight subjective contours (SSC), using an a contrario approach to control the false detection rate. The algorithm exploits the tips of line segments produced by the well-known parameter-less LSD method. The subjective straight contours are obtained by grouping free tips of parallel line sets, together with aligned short edge pieces. This detection is fully automatic and is demonstrated on a set of images containing subjective contours.
Double phase retrieval algorithm
The problem of compressive imaging is addressed using natural randomization by means of a multiply scattering medium. To utilize the medium in this way, its corresponding transmission matrix must be estimated. For calibration purposes, we use a digital micromirror device (DMD) as a simple, cheap, and high-resolution binary intensity modulator. We propose a phase retrieval algorithm which is well adapted to intensity-only measurements on the camera, and to the input binary intensity patterns, both to estimate the complex transmission matrix as well as image reconstruction. We demonstrate promising experimental results for the proposed double phase retrieval algorithm using the MNIST dataset of handwritten digits as example images.
Robust phase retrieval with the prSAMP algorithm
In phase retrieval, the goal is to recover a complex signal from the magnitude of its linear measurements. While many well-known algorithms guarantee deterministic recovery of the unknown signal using i.i.d. random measurement matrices, they suffer serious convergence issues for some ill-conditioned measurement matrices. As an example, this happens in optical imagers using binary intensity-only spatial light modulators to shape the input wavefront. The problem of ill-conditioned measurement matrices has also been a topic of interest for compressed sensing researchers during the past decade. Using recent advances in generic compressed sensing, we propose a new phase retrieval algorithm that well-behaves for a large class of measurement matrices, including Gaussian and Bernoulli binary i.i.d. random matrices, using both sparse and dense input signals. This algorithm is also robust to the strong noise levels found in some imaging applications.
Intensity-only measurement of partially uncontrollable transmission matrix
Transmission matrices (TMs) have become a powerful and widely used tool to describe and control wave propagation in complex media. In certain scenarios the TM is partially uncontrollable, complicating its identification and use. In standard optical wavefront shaping experiments, uncontrollable reflections or imperfect illumination may be the cause; in reverberating cavities, uncontrollable reflections off the walls have that effect. Here we employ phase retrieval techniques to identify such a partially uncontrollable TM solely based on random intensity-only reference measurements. We demonstrate the feasibility of our method by focusing both on a single target as well as on multiple targets in a microwave cavity, using a phase-binary Spatial-Microwave-Modulator.
Gestaltic grouping of line segments
Using simple grouping rules in Gestalt theory, one may detect higher level features (geometric structures) in an image from elementary features. By recursive grouping of already detected geometric structures a bottom-up pyramid could be built that extracts increasingly complex geometric features from the input image. Taking advantage of the (recent) advances in reliable line segment detectors, we propose three feature detectors along with their corresponding detailed algorithms that constitute one step up in this pyramid. For any digital image, our unsupervised algorithm computes three classic Gestalts from the set of pre-detected line segments: good continuations, non-local alignments, and bars. The methodology is based on a common stochastic a contrario model yielding three simple detection formulas, characterized by their number of false alarms. This detection algorithm is illustrated on several digital images.
Handwritten signature identification and verification
A new method for signature identification and verification based on contourlet transform (CT) is proposed. This method uses contourlet coefficient as the feature extractor and Support Vector Machine (SVM) as the classifier. In proposed method, first signature image is normalized based on size. After preprocessing, contourlet coefficients are computed on specified scale and direction. Next, all extracted coefficients are fed to a layer of SVM classifiers as feature vector. The number of SVM classifiers is equal to the number of classes. Each SVM classifier determines if the input image belongs to the corresponding class or not. The main characteristic of proposed method is independency to nation of signers. Two experiments on two signature sets are performed.
Fast Phase Retrieval for High Dimensions
Here, we address fundamental scaling issues that hinder phase retrieval (PR) in high dimensions. We show that, if the measurement matrix can be put into a generalized block-diagonal form, a large PR problem can be solved on separate blocks, at the cost of a few extra global measurements to merge the partial results. We illustrate this principle using two distinct PR methods, and discuss different design trade-offs. Experimental results indicate that this block-based PR framework can reduce computational cost and memory requirements by several orders of magnitude.
An Analysis and Improvement of the BLS-GSM Denoising
Modeling image properties using the Gaussian scale mixture (GSM) model in a multiresolution transform space is the basic idea of a denoising algorithm proposed by Portilla et al. Under this model and using the correlations between pyramid coefficients, the Bayesian least squares (BLS) of each coefficient is used to estimate its original value. Here, we analyze and discuss the BLS-GSM algorithm, its drawbacks and benefits in more detail. An analytical parameter study of this denoising approach is provided as well. Additionally, we propose a localized version of this algorithm and experimentally show that it outperforms the original method both numerically and visually. We also show that the resulting method is state-of-the-art in terms of PSNR.
Building software and hardware of some kinds of racing robots such as line tracking, micromouse, etc.
Robotic Arm Controller
A computer program to control a 3-axis robotic arm.
Project Management software
After comprehensive studies performed by the product owners on similar apps e.g. Bitrix24, Jira and Trello, we started development based on agile methodology. In the beginning, we customized Fuse Angular template – a material design admin template with Angular 7 – taking the key role of client app skeleton. Then, we implemented a client engine to handle REST requests, Web-Socket communications, authentication, accessibility, client paging and sorting, etc. On the back-end side, we used .NET Core 2.2 to implement authentication, permission system, Web-Socket, etc.
Behavioral pattern mining module
The analysis module of performance behavior patterns compares the performance of all organizational units (including employees, evaluators, and organizations) with the reference model, using effective indicators in the evaluation process. The reference model is created by the user, who selects relevant individuals or organizations, and then the degree of behavioral similarity of people with the reference model is analyzed.
·Creating a behavioral reference model,
·Comparing performance and calculating the degree of behavioral similarity with the reference model,
·Implementing the analysis of behavioral patterns in various processes, such as employee behavior, organizational behavior, financial behavior, etc.
Frequent itemset mining module
The module for identifying frequent patterns and dependency rules, which utilizes machine learning algorithms, explores relationships and interdependencies within data. Such rules have applications in various fields, such as analyzing customers’ shopping carts, identifying patterns between products and services, examining how different indicators influence each other, etc.
• Another feature of this module is the ability to clean data using different data mining templates.
• The ability to determine the type of impact between features, whether it is positive, negative, or non-impact, by defining the percentage of confidence and support.
• Identification of “if… then…” rules.
• Data cleaning
Database administration, SQL tuning, and refactoring a few subsystems of Click accounting system to decrease database complexity and to boost performance in heavy reports. To fulfill our desire, we used our knowledge about execution plans of SQL server and its outstanding features e.g. inline functions, recursive CTEs, semi-structured data, dynamic queries, etc. We worked on other subsystems as well: finance, treasury, trading, and salary.
React, Redux, Material UI, and Tailwind CSS technologies are employed to implement 3 subsystems: Content Management (CMS), Automation, and Online Ordering System for Fuel Products. We used .NET 5 for the server side. The client side applies Fuse – a robust and lightweight React-Redux template – to accelerate implementation speed.
Online Hotel Reservation System
A web application for hotels to record the status of their rooms, services and rates
A web application for travel agencies to sale the availability of hotels with different models as well as selling flight tickets
A public web application for people and travelers to reserve available hotel rooms.
Authentication and Authorization System as a Service
Payment as a Service
A web service that can facilitate payment processes for applications.
Automatic Parking Gate
A computer program that connects to a specially designed piece of hardware that automatically opens the entrance gate of a complex or building for a specific list of vehicles.