Projects from my CPSC 425 computer vision class.
Implementations are done with minimal use of higher-level computer vision or ML libraries.

Image Filtering and Hybrid Images
Scaled Representations, Face Detection, Image Blending
Texture Synthesis
RANSAC and Panorama Stiching
Scene Recognition with Bag of Words
Deep Learning

Image Filtering and Hybrid Images

Creating Hybrid images using open-cv and numpy.

Scaled Representations, Face Detection, Image Blending

Face Detection

Face detection with template matching using open-cv and numpy.

Image Blending

Image blending using laplacian and gaussian pyramids using open-cv and numpy.

Texture Synthesis

Wrote the a texture synthesis function based on Efros and Leung.

RANSAC and Panorama Stiching

Wrote functions for panorama stiching with RANSAC and SIFT using numpy.

Scene Recognition with Bag of Words

Wrote functions for a KNN and SVM classifiers with numpy and sklearn.

Deep Learning

Image Classification

Tuned the hyperparameters of an CNN-based image classification.

Net(
  (cnn): Sequential(
    (0): Conv2d(3, 16, kernel_size=(3, 3), stride=(1, 1))
    (1): ReLU()
    (2): MaxPool2d(kernel_size=2, stride=2, padding=0, dilation=1, ceil_mode=False)
    (3): Conv2d(16, 16, kernel_size=(3, 3), stride=(1, 1))
    (4): ReLU()
    (5): MaxPool2d(kernel_size=2, stride=2, padding=0, dilation=1, ceil_mode=False)
    (6): Conv2d(16, 64, kernel_size=(3, 3), stride=(1, 1))
    (7): ReLU()
  )
  (linear_layers): MLP(
    (model): ModuleList(
      (0): Linear(in_features=1024, out_features=16, bias=True)
      (1): ReLU()
      (2): Linear(in_features=16, out_features=16, bias=True)
      (3): ReLU()
      (4): Linear(in_features=16, out_features=10, bias=True)
    )
  )
)

Image Segmentation

Experimented on the types of inputs Mask R-CNN will succeed and fail on.

MaskRCNN(
  (transform): GeneralizedRCNNTransform(
      Normalize(mean=[0.485, 0.456, 0.406], std=[0.229, 0.224, 0.225])
      Resize(min_size=(800,), max_size=1333, mode='bilinear')
  )
  (backbone): BackboneWithFPN(
    (body): IntermediateLayerGetter(
      (conv1): Conv2d(3, 64, kernel_size=(7, 7), stride=(2, 2), padding=(3, 3), bias=False)
      (bn1): FrozenBatchNorm2d(64, eps=0.0)
      (relu): ReLU(inplace=True)
      (maxpool): MaxPool2d(kernel_size=3, stride=2, padding=1, dilation=1, ceil_mode=False)
      (layer1): Sequential(
        (0): Bottleneck(
          (conv1): Conv2d(64, 64, kernel_size=(1, 1), stride=(1, 1), bias=False)
          (bn1): FrozenBatchNorm2d(64, eps=0.0)
          (conv2): Conv2d(64, 64, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
          (bn2): FrozenBatchNorm2d(64, eps=0.0)
          (conv3): Conv2d(64, 256, kernel_size=(1, 1), stride=(1, 1), bias=False)
          (bn3): FrozenBatchNorm2d(256, eps=0.0)
          (relu): ReLU(inplace=True)
          (downsample): Sequential(
            (0): Conv2d(64, 256, kernel_size=(1, 1), stride=(1, 1), bias=False)
            (1): FrozenBatchNorm2d(256, eps=0.0)
          )
        )
...
      (relu): ReLU(inplace=True)
      (mask_fcn_logits): Conv2d(256, 91, kernel_size=(1, 1), stride=(1, 1))
    )
  )
)