Source author record

Mohammad Jafari

Mohammad Jafari appears in the imported research catalog. Authorship, coauthor and topic links are available while profile ownership is still unclaimed.

ResearcherUnclaimed source record

Computer Vision Artificial Intelligence Distributed, Parallel, and Cluster Computing Machine Learning Robotics

Catalog footprint

What is connected

4works

5topics

4close collaborators

Actions

Connect this record

Open graph Browse works

Inspect adjacent papers, topics, institutions and collaborators without losing the researcher page.

Building this map preview

BZPEER is loading the nearby papers, people, topics and institutions for this page.

preprint2022arXiv

HealNet -- Self-Supervised Acute Wound Heal-Stage Classification

Identifying, tracking, and predicting wound heal-stage progression is a fundamental task towards proper diagnosis, effective treatment, facilitating healing, and reducing pain. Traditionally, a medical expert might observe a wound to determine the current healing state and recommend treatment. However, sourcing experts who can produce such a diagnosis solely from visual indicators can be difficult, time-consuming and expensive. In addition, lesions may take several weeks to undergo the healing process, demanding resources to monitor and diagnose continually. Automating this task can be challenging; datasets that follow wound progression from onset to maturation are small, rare, and often collected without computer vision in mind. To tackle these challenges, we introduce a self-supervised learning scheme composed of (a) learning embeddings of wound's temporal dynamics, (b) clustering for automatic stage discovery, and (c) fine-tuned classification. The proposed self-supervised and flexible learning framework is biologically inspired and trained on a small dataset with zero human labeling. The HealNet framework achieved high pre-text and downstream classification accuracy; when evaluated on held-out test data, HealNet achieved 97.7% pre-text accuracy and 90.62% heal-stage classification accuracy.

preprint2021arXiv

Reciprocal Landmark Detection and Tracking with Extremely Few Annotations

Localization of anatomical landmarks to perform two-dimensional measurements in echocardiography is part of routine clinical workflow in cardiac disease diagnosis. Automatic localization of those landmarks is highly desirable to improve workflow and reduce interobserver variability. Training a machine learning framework to perform such localization is hindered given the sparse nature of gold standard labels; only few percent of cardiac cine series frames are normally manually labeled for clinical use. In this paper, we propose a new end-to-end reciprocal detection and tracking model that is specifically designed to handle the sparse nature of echocardiography labels. The model is trained using few annotated frames across the entire cardiac cine sequence to generate consistent detection and tracking of landmarks, and an adversarial training for the model is proposed to take advantage of these annotated frames. The superiority of the proposed reciprocal model is demonstrated using a series of experiments.

preprint2015arXiv

An efficient parallel algorithm for computing determinant of non square matrices

One of the most significant challenges in Computing Determinant of Rectangular Matrices is high time complexity of its algorithm. Among all definitions of determinant of rectangular matrices, used definition has special features which make it more notable. But in this definition, C(n m) sub matrices of the order m*m needed to be generated that put this problem in NP hard class. On the other hand, any row or column reduction operation may hardly lead to diminish the volume of calculation. Therefore, in this paper we try to present the parallel algorithm which can decrease the time complexity of computing the determinant of non-square matrices to O(pow(n,2)).

preprint2014arXiv

Applied Neural Cross-Correlation into the Curved Trajectory Detection Process for Braitenberg Vehicles

Curved Trajectory Detection (CTD) process could be considered among high-level planned capabilities for cognitive agents, has which been acquired under aegis of embedded artificial spiking neuronal circuits. In this paper, hard-wired implementation of the cross-correlation, as the most common comparison-driven scheme for both natural and artificial bionic constructions named Depth Detection Module(DDM), has been taken into account. It is manifestation of efficient handling upon epileptic seizures due to application of both excitatory and inhibitory connections within the circuit structure. Presented traditional analytic approach of the cross-correlation computation with regard to our neural mapping technique and the acquired traced precision have been turned into account for coherent accomplishments of the aforementioned design in perspective of the desired accuracy upon high-level cognitive reactions. Furthermore, the proposed circuit could be fitted into the scalable neuronal network of the CTD, properly. Simulated denouements have been captured based on the computational model of PIONEER mobile robot to verify characteristics of the module, in detail.