🧠 Computer Vision Model Collection

다양한 컴퓨터 비전 모델을 실험하고 결과를 정리한 저장소입니다.

---

📚 Table of Contents

• 1. Image Classification
  • 1.1 ConvNet
  • 1.2 ResNet50 Finetuning
  • 1.3 Vision Transformer (ViT)
• 2. Video Classification: S3D
• 3. Object Detection: Faster-RCNN
• 4. Image Translation: CycleGAN
• 5. Metric Learning: Triplet Loss

---

1. Image Classification

1.1 ConvNet

• Dataset: CIFAR-10
• 🧩 Model: Custom ConvNet

📊 Results

---

1.2 ResNet50 Finetuning

• Dataset: CIFAR-10

🧩 Model: ResNet50 (Fine-tuning)

기존 ResNet50 모델을 불러와 CIFAR-10에 맞게 fine-tuning 진행

---

1.3 Vision Transformer (ViT)

• Dataset: CIFAR-10

🧩 Model: Vision Transformer

📊 Results

아직 더 학습이 필요하나, 제한된 GPU에서 최대로 학습한 결과:

---

2. Video Classification: S3D

• Dataset: UCF-11

🧩 Model: S3D

📊 Results

 

download.mp4

---

3. Object Detection: Faster-RCNN

• Dataset: VOC2007

🧩 Model: Faster-RCNN ([Paper](https://arxiv.org/abs/1506.01497))

📊 Results

---

4. Image Translation: CycleGAN

• Dataset: horse2zebra

🧩 Model: CycleGAN ([Paper](https://arxiv.org/abs/1703.10593))

📊 Results

아직 더 학습이 필요하나, 최대한으로 생성된 결과:

---

5. Metric Learning: Triplet Loss

• Dataset: MNIST
• Task: 숫자 임베딩 학습
• Method: Triplet Loss

🧩 모델 및 임베딩 시각화

📊 결과