پروژه رشته ict با یونیتی (7955)
پنج سال پیش منتشر شده
تعداد بازدید: 351
کد پروژه: 175726
شرح پروژه
سلام وقت بخیر
توضیوح پروژه رو میفرستم لطفا بررسی کنید و اطلاع بدین
Project A.1
2020 Visualization Challenge: physical properties for mechanical
models
The goal of this challenge is to visualize in a Unity application some
physical properties (pressure, aerodinamics) for a 3D mechanical model (a
car). Datasets is made of a 3D model and a list of possible data localized in
the 3D space around (and on the model.
Data includes pressure values (localized on
the surface of the 3D model) and
ow
lines (sequences of 3D points - localized in
space/time associated to air). See an example
in the gure aside.
Students may design di erent interfaces and visualization (static or dy-
namic). Visualization strategies can include color mapping, 3D solid ani-
mations, particle systems. Projects will be presented at the nal exam and
ranked.
Dataset is courtesy of Altair www.altair.com.
Project B.1
Feature matching for Structure-from-Motion
Using the SfM software COLMAP https://colmap.github.io/, tests
the e ectiveness of di erent local descriptors in the 3D reconstruction (mea-
suring the number of points, the average track length and the reprojection
error). You need to provide to the softeware the list of keypoint positions
and possible matchings.
In the experiments, you can use the descriptors
SURF (implemented in OpenCV/MATLAB)
ORB (implemented in OpenCV/MATLAB)
BRIEF (implemented in OpenCV/MATLAB)
binGAN (implemented in Python https://github.com/maciejzieba/
binGAN)
TFeat (in Python https://github.com/vbalnt/tfeat)
LIFT (code in Python/C https://github.com/cvlab-epfl/LIFT).
Performances must be evaluated on the datasets Fountain and Castle (https:
//github.com/openMVG/SfM_quality_evaluation/tree/master/Benchmarking_
Camera_Calibration_2008)and Tiso and Portello (http://www.dei.unipd.
it/~sim1mil/materiale/3Drecon/).
Project B.2
Local feature compression using autoencoders - SfM tests
Design a compression strategy for local SURF descriptors using autoen-
coders. Training data can be generated using the images of dataset Portello
and Castle. Testing must be done on dataset FountainP-11 and Tiso (avail-
able at https://github.com/openMVG/SfM_quality_evaluation/tree/master/
Benchmarking_Camera_Calibration_2008 and http://www.dei.unipd.it/
~sim1mil/materiale/3Drecon/). Software must be implemented in MAT-
LAB, Keras or Pytorch.
Testing on 3D reconstruction using SfM The reconstructed de-
scriptors (only for the test set) are used to perform a SfM reconstruction
using COLMAP (using the two test dataset).
من یکی از این پروژه ها رو نیاز دارم, لطفا بفرمایید کدوم میتونن با وقت کمتری نوشته بشن
رشته ارشد Ict
مهارت ها و تخصص های مورد نیاز
مهلت برای انجام
10روز
وضعیت مناقصه
بسته
درباره کارفرما
عضویت سیزده سال پیش
