Additional Course | Convolutional Neural Networks – Examples and Best Practices

Given by University Politehnica of Bucharest

Additional Webinar Course on Convolutional Neural Networks – Examples and Best Practices

25 October 2018 at 10:00 AM Peru time (17:00 Italy time and 18:00 Romania time), Liviu Stefan and Gabriel Constantin (UPB) will conduct a webinar on Convolutional Neural Networks – Examples and Best Practices.

It will focus on presenting some practical aspects of Convolutional Neural Networks. Firstly, it is going to explore the architecture of MobileNet, a small yet powerful Convolutional Neural Network, addressing the core layers and exploring the trade-off between the number of parameters and accuracy. Secondly, it is going to present a guide to several methods for improving the training results and output accuracy of Convolutional Neural Networks, exploring concepts regarding network tuning, data augmentation and other interesting methods. Several practical code examples will be analyzed.

TRACK A: Mobile and Computer Vision

MOBILE PROGRAMMING 1  (basic) – 26-29 JUNE 2017
Course objective: Creating mobile web apps
Course contents: Sample development using HTML5-Javascript, PHP; Hybrid apps: translation to native code via Cordova

MOBILE PROGRAMMING  2  (adv) – 10,11,12 and 14 JULY 2017
Course objective: Creating mobile native Android apps
Course contents: Sample development using native code via Android Studio; Android API; JNI Java Native Interface; bridging Open CV using JNI

COMPUTER VISION (adv. onsite course IN FLORENCE) – 15-19 NOVEMBER 2017
Course objective: Creating computer vision recognition applications
Course contents: OPENCV framework, Lib SVM and CNN Frameworks; SIFT detector and descriptors; CNN models and descriptors; SVM Classifiers, CNN detectors and classifiers. Sample application development
Continue reading “TRACK A: Mobile and Computer Vision”

TRACK B: Multimedia big data

PRE-PROCESSING OF VISUAL INFORMATION  (basic) – 17-18 JULY 2017
Course objective: Familiarizing with image processing
Course contents: Image/video content pre-processing: contrast enhancement, smoothing and noise removal techniques

MULTIMEDIA CONTENT REPRESENTATION  (adv) – 24-27 JULY 2017
Course objective: Learning multimedia features and their representation
Course contents: Data representation, data analysis techniques; content description including color, texture, shape, features, motion, temporal structure, audio, text; social data and common analytics techniques; data normalization and decorrelation techniques

MULTIMEDIA CLASSIFICATION (adv) – 7-8 SEPTEMBER 2017
Course objective: Learning tools for categorizing multimedia contents
Course contents: Unsupervised clustering: hierarchical clustering, k-means; supervised classification: k-NN, Support Vector Machines, decision trees for information retrieval and content understanding

MULTIMEDIA RETRIEVAL   (adv – onsite course IN BUCAREST) – 7-8 and 11-12 SEPTEMBER 2017
Course objective: Learning principles of content-based retrieval from large datasets
Course contents: Information retrieval systems; data indexing data similarity; data fusion; relevance feedback; performance metrics

Continue reading “TRACK B: Multimedia big data”

TRACK C: 3D

Given by IMT Lille Douai

3D scanning (basic) – 22-23 JUNE 2017
Course objective: Introduction to fundamentals of 3D acquisition
Course contents: 3D scanning; stereo vision, triangulation, laser scanning, structured light; time-of-flight cameras

3D scan pre-processing  (adv) – 3-5 JULY 2017
Course objective: Learning 3D pre-processing operations
Course contents: Triangular meshes; geometry-processing: pipeline based on triangular meshes; geometry processing: mesh parameterisation and mesh quantities; MeshLab

3D SCAN MATCHING AND REGISTRATION (basic) – 12-13 SEPTEMBER 2017
Course objective: Learning tools for Augmented Reality applications
Course contents: Basic concepts of Augmented Reality

3D SCAN MATCHING AND REGISTRATION (adv, onsite course IN LILLE) – 13-16 NOVEMBER 2017
Course objective: Learning principles for matching 3D data
Course contents: Matching algorithms; Iterative Closest Point algorithm, for the alignment of the cloud points; Point Cloud Library

Continue reading “TRACK C: 3D”