CV project workflow in Computer Vision - Cheat Sheet & Quick Revision

Recall & Review

beginner

What is the first step in a computer vision project workflow?

The first step is to clearly define the problem you want to solve, such as object detection, image classification, or segmentation.

beginner

Why is data collection important in a CV project?

Data collection provides the images or videos needed to train and test the model. Good quality and diverse data help the model learn better.

intermediate

What does data preprocessing involve in a CV workflow?

Data preprocessing includes resizing images, normalizing pixel values, augmenting data, and labeling images to prepare them for training.

beginner

What is the purpose of model training in a CV project?

Model training teaches the computer vision model to recognize patterns in the data by adjusting its parameters to minimize errors.

intermediate

How do you evaluate a computer vision model's performance?

You evaluate it using metrics like accuracy, precision, recall, or IoU (Intersection over Union) depending on the task, using a separate test dataset.

What is the main goal of data augmentation in CV projects?

A Label the images automatically

B Reduce the size of the dataset

C Remove noisy images

D Increase the size and diversity of the training data

Data augmentation creates new training examples by modifying existing images, helping the model generalize better.

Which step comes directly after model training in a typical CV workflow?

A Data collection

B Data preprocessing

C Model evaluation

D Problem definition

After training, the model is evaluated to check how well it performs on unseen data.

Why do we split data into training and test sets?

A To evaluate model performance on unseen data

B To train two different models

C To reduce data size

D To label data automatically

Splitting data helps test if the model can generalize beyond the data it learned from.

Which metric is commonly used for object detection tasks?

A Accuracy

B IoU (Intersection over Union)

C Mean Squared Error

D Perplexity

IoU measures how well predicted bounding boxes overlap with ground truth boxes in object detection.

What is the role of labeling in a CV project?

A To assign correct answers to images for supervised learning

B To organize files

C To compress images

D To increase image resolution

Labeling provides the correct answers the model learns to predict during training.

Describe the main steps in a computer vision project workflow from start to finish.

Think about what you do first, how you prepare data, then how you teach and test the model.

Explain why data preprocessing and augmentation are important before training a CV model.

Consider how raw images might vary and how the model benefits from more varied examples.