Intermediate everything in AI & ML, page 7
Why averaging shots reduces image noise
Random noise averages out while signal stays, so SNR rises with the square root of frame count; limits include motion and fixed-pattern noise.
The role of anchor boxes in detectors
Anchors are reference boxes of set scales and ratios; the model regresses offsets, framing detection as classification plus refinement; downsides are tuning, imbalance.
Triangulating 3D points from two views
Back-project each 2D point into a ray, intersect rays from both cameras using known poses, solve via DLT and refine by minimizing reprojection error.
Single vs multi-image super-resolution
SISR hallucinates detail from one image, multi-image fuses sub-pixel shifts; pixel losses blur, perceptual or GAN losses add realism.
Non-maximum suppression in detection
Detectors emit many overlapping boxes per object; NMS keeps the highest-scoring box and removes others above an IoU threshold.
One-stage vs two-stage detectors
One-stage predicts boxes directly for speed; two-stage proposes then refines for accuracy; focal loss narrows the gap.
Designing a multi-object tracker
Detect per frame, predict motion with a filter, associate via IoU and appearance, manage track lifecycle, handle occlusion with re-ID.
Challenges deploying a model on edge hardware
Limited memory and compute cause latency, thermal and power limits, accuracy loss from compression, operator support gaps.
When a homography is a valid model
Homography holds for pure rotation or a planar scene; it fails with translation plus 3D parallax, where epipolar geometry applies.
Contrastive learning vs masked image modeling
Contrastive aligns augmented views via instance discrimination; MAE reconstructs masked patches; they differ in augmentation and fine-tuning.
Zero-shot classification with CLIP
Encode image and label prompts into a shared space, compare via cosine similarity, pick the highest.
Point cloud vs voxel grid vs NeRF
Point clouds are sparse and fast but unstructured, voxels are regular for collision checks but memory-heavy, NeRFs render photorealistically but are slow.
Cross-attention in transformer VQA models
Text queries attend over image regions, learning alignment that grounds words to visual content.
The data association problem in SLAM
Matching observations to landmarks, why wrong matches corrupt the map, robust techniques like RANSAC and descriptor matching.
Core components of visual SLAM
Tracking estimates per-frame pose, mapping builds and refines the 3D map, loop closure detects revisits and corrects drift.
Adapting a 2D CNN for video action recognition
Run the 2D CNN per frame, pool features over time, optionally add two-stream or 3D conv.
Deploying segmentation on edge devices
Pick efficient architectures, apply INT8 quantization, distill from a large teacher.
Improving small object detection
Raise input resolution and tile, use feature pyramids for high-res features, tune anchors and copy-paste augmentation.
What is Bundle Adjustment and why is it tractable?
Jointly refine 3D points and camera poses by minimizing reprojection error, expensive due to many coupled parameters; sparsity of the Jacobian and the Schur complement make it tractable.
Why learn detection and description jointly like SuperPoint?
A shared backbone jointly optimizes detection and description for matching, sharing computation and learning data-driven robustness instead of hand-crafted heuristics.
We are hiring for this. Every open role lists the topics its interview covers, so you can prepare for the real thing rather than guessing.
See open roles