Skip to content

umutsaydam/ColorAnalysisByObjects

Repository files navigation

Color Analysis By Objects

Note

The app is still in beta and There is a stage to finish (identify data by classifying because accurate detection rate will increase). If you use this app, you will need the database. I can share with you because I gues I will not deal with this project anymore. Contact me umutsaydam24@gmail.com. Before starting this project, I had no idea how to do this, but when I started the project, everything became clear. The make a long story, Just Do It !

Amaç

Bu proje ile obje bazında en çok kullanılan üç rengin belirlenerek veri tabanına kaydedilmesi ve böylece dönemlere göre insanların tercih ettiği renk alışkanlıklarının ortaya çıkarılması amaçlanmaktadır.

🦶 Gerçekleştirilen-adımlar

1. Makineyi, kategorileştirilmiş veriler ile eğitmek

  • Kategorilerine göre (giyim vs.) sınıflandırılmış verileri ayrı ayrı eğiterek nesneyi doğru tespit etme oranı artırıldı.

2. Instance segmentation Region of interest

  • Instance segmentation ile ilgili nesnenin tam konumlarını tespit edip, Region of interest algoritmasıyla tespit edilen nesnenin koordinatlarına göre kırpılması sağlandı.

stages_of_detection

3. Ortalama renk analizi ve Kmeans

  • Kırpılmış nesneler üzerinde ağırlıklı olarak bulunan üç renk tespit edilerek verilerin kategoriye göre depolanması.

    result_of_detection

4. Depolanan veriler doğrultusunda kullanıcıların kategoriye ve zamana göre renk alışkanlıklarının tespiti

  • Toplanan veriler sayesinde network grafiği ile kategorilere göre renk alışkanlıklarının tespiti.

graph

statistics

5. Kurulum

  • requirements.txt içerisindekiler kurulmalı.
    pip install -r FILE_PATH/requirements.txt
    

Aim

With this project, it is aimed to determine the three most used colors on the basis of objects and to save them in the database, thus revealing the color habits preferred by people according to the periods.

🦶 Steps-Taken

1. Makineyi, kategorileştirilmiş veriler ile eğitmek

  • Kategorilerine göre (giyim vs.) sınıflandırılmış verileri ayrı ayrı eğiterek nesneyi doğru tespit etme oranı artırıldı.

2. Instance segmentation Region of interest

  • Instance segmentation ile ilgili nesnenin tam konumlarını tespit edip, Region of interest algoritmasıyla tespit edilen nesnenin koordinatlarına göre kırpılması sağlandı.

stages_of_detection

3. Ortalama renk analizi ve Kmeans

  • Kırpılmış nesneler üzerinde ağırlıklı olarak bulunan üç renk tespit edilerek verilerin kategoriye gmre depolanması.

    result_of_detection

4. Depolanan veriler doğrultusunda kullanıcıların kategoriye ve zamana göre renk alışkanlıklarının tespiti

  • Toplanan veriler sayesinde network grafiği ile kategorilere göre renk alışkanlıklarının tespiti.

graph

statistics

5. Installation

  • requirements.txt contents must be installed.
    pip install -r FILE_PATH/requirements.txt