Learn/Course 10
Deep learning · weeks 13

Image ML

A factory has few positive examples of costly visual defects.

Opening production brief

A measurable problem, before a model.

Users

Operators, product owners, and people affected by the vision decision.

Data

NCHW tensors, convolution, receptive fields, augmentation with version, owner, event time, and availability contracts.

Constraints

A factory has few positive examples of costly visual defects. Every solution must state latency, cost, capacity, and fallback limits.

Success

Offline evidence plus a deployable resolution: CNN and transfer-learning baselines with augmentation, calibration, and monitoring.

CHAPTER 1

Frame and baseline

Turn the production problem into explicit data, metric, and baseline contracts.

  1. 01

    Images are data contracts

    Use images are data contracts to move the vision production brief toward a defensible release.

    65–90 min · checkpoint · deterministic lab
  2. 02

    Convolution under the microscope

    Use convolution under the microscope to move the vision production brief toward a defensible release.

    65–90 min · checkpoint · deterministic lab
CHAPTER 2

Build and stress

Formalize the machinery, test counterexamples, and expose system limits.

  1. 03

    Receptive fields, normalization, and residuals

    Use receptive fields, normalization, and residuals to move the vision production brief toward a defensible release.

    65–90 min · checkpoint · deterministic lab
  2. 04

    Transfer learning for classification, detection, and segmentation

    Use transfer learning for classification, detection, and segmentation to move the vision production brief toward a defensible release.

    65–90 min · checkpoint · deterministic lab
CHAPTER 3

Resolve and operate

Join model behavior to architecture, observability, rollout, and rollback.

  1. 05

    Vision in the wild: thresholds, latency, and drift

    Use vision in the wild: thresholds, latency, and drift to move the vision production brief toward a defensible release.

    65–90 min · checkpoint · deterministic lab
  2. HW

    Rare-defect vision system

    Design grouped evaluation, transfer learning, calibration, batch-one export, and camera-drift monitoring.

    Rubric · staged hints · reference resolution
Prerequisite guidance

1 concepts make this faster.

Prerequisites are guidance, never hard gates.

C09Neural networksA small network built from forward pass through backpropagation and optimization.
Ready to work the problem?

Images are data contracts

Start lesson 01