Car damage detection using azure machine learning or azure artificial intelligence

Surya Narayana Korivipadu 21 Reputation points
2022-01-24T10:41:45.263+00:00

Hi. Can someone please guide me how to detect damages in the car from car images using Azure Machine Learning or Azure AI?

I'm planning to use image classification computer vision solution as a first step to classify if car is damaged or not, then as a second step use object detection to identify which parts of the car are damaged.

I'm a beginner in AI and ML. Am I going with the correct approach or is there any other way to solve my problem?

Azure Machine Learning
Azure Machine Learning
An Azure machine learning service for building and deploying models.
2,576 questions
Azure Computer Vision
Azure Computer Vision
An Azure artificial intelligence service that analyzes content in images and video.
313 questions
0 comments No comments
{count} votes

2 answers

Sort by: Most helpful
  1. GiftA-MSFT 11,151 Reputation points
    2022-01-25T03:33:41.933+00:00

    Hi, I think you're heading in the right direction. Since this is a classification problem, you'd want to use Azure ML (computer vision) to classify the images. To identify which parts have been damaged, Custom Vision's object detector seems to be a viable solution. Please refer to the documentation links provided.

    1 person found this answer helpful.

  2. Alvaro Mathis 0 Reputation points
    2023-10-25T16:01:18.56+00:00
    Um Autoschäden mithilfe von Azure Machine Learning oder Azure Artificial Intelligence (KI) zu erkennen, muss ein Modell erstellt werden, das Bilder von Autos analysieren und feststellen kann, ob sie beschädigt wurden. Hier ist ein allgemeiner Überblick über die Schritte, die zum Aufbau eines solchen Systems erforderlich sind:
    
    Datensammlung:
    
    Sammeln Sie einen Datensatz mit Fahrzeugbildern, von denen einige Schäden aufweisen, andere jedoch nicht.
    Kommentieren Sie die Bilder, um anzugeben, ob sie Schäden enthalten, und wenn ja, um welche Art und Ausmaß es sich handelt.
    Datenvorverarbeitung:
    
    Ändern Sie die Größe der Bilder und standardisieren Sie sie auf ein einheitliches Format.
    Teilen Sie den Datensatz in Trainings-, Validierungs- und Testsätze auf.
    Modellauswahl:
    
    Wählen Sie ein geeignetes Deep-Learning-Modell für die Bildklassifizierung. Zu den beliebten Optionen gehören Convolutional Neural Networks (CNNs).
    
    [autoankauf](https://www.bares4cars.de/)
    
    0 comments No comments