Picture for Anand Bhattad

Anand Bhattad

Latent Intrinsics Emerge from Training to Relight

May 31, 2024
Viaarxiv icon

Videoshop: Localized Semantic Video Editing with Noise-Extrapolated Diffusion Inversion

Add code
Mar 22, 2024
Figure 1 for Videoshop: Localized Semantic Video Editing with Noise-Extrapolated Diffusion Inversion
Figure 2 for Videoshop: Localized Semantic Video Editing with Noise-Extrapolated Diffusion Inversion
Figure 3 for Videoshop: Localized Semantic Video Editing with Noise-Extrapolated Diffusion Inversion
Figure 4 for Videoshop: Localized Semantic Video Editing with Noise-Extrapolated Diffusion Inversion
Viaarxiv icon

Generative Models: What do they know? Do they know things? Let's find out!

Add code
Nov 28, 2023
Viaarxiv icon

Shadows Don't Lie and Lines Can't Bend! Generative Models don't know Projective Geometry...for now

Add code
Nov 28, 2023
Viaarxiv icon

Improving Equivariance in State-of-the-Art Supervised Depth and Normal Predictors

Add code
Sep 28, 2023
Figure 1 for Improving Equivariance in State-of-the-Art Supervised Depth and Normal Predictors
Figure 2 for Improving Equivariance in State-of-the-Art Supervised Depth and Normal Predictors
Figure 3 for Improving Equivariance in State-of-the-Art Supervised Depth and Normal Predictors
Figure 4 for Improving Equivariance in State-of-the-Art Supervised Depth and Normal Predictors
Viaarxiv icon

OBJECT 3DIT: Language-guided 3D-aware Image Editing

Jul 20, 2023
Figure 1 for OBJECT 3DIT: Language-guided 3D-aware Image Editing
Figure 2 for OBJECT 3DIT: Language-guided 3D-aware Image Editing
Figure 3 for OBJECT 3DIT: Language-guided 3D-aware Image Editing
Figure 4 for OBJECT 3DIT: Language-guided 3D-aware Image Editing
Viaarxiv icon

Blocks2World: Controlling Realistic Scenes with Editable Primitives

Add code
Jul 13, 2023
Figure 1 for Blocks2World: Controlling Realistic Scenes with Editable Primitives
Figure 2 for Blocks2World: Controlling Realistic Scenes with Editable Primitives
Figure 3 for Blocks2World: Controlling Realistic Scenes with Editable Primitives
Figure 4 for Blocks2World: Controlling Realistic Scenes with Editable Primitives
Viaarxiv icon

UrbanIR: Large-Scale Urban Scene Inverse Rendering from a Single Video

Add code
Jun 16, 2023
Figure 1 for UrbanIR: Large-Scale Urban Scene Inverse Rendering from a Single Video
Figure 2 for UrbanIR: Large-Scale Urban Scene Inverse Rendering from a Single Video
Figure 3 for UrbanIR: Large-Scale Urban Scene Inverse Rendering from a Single Video
Figure 4 for UrbanIR: Large-Scale Urban Scene Inverse Rendering from a Single Video
Viaarxiv icon

StyleGAN knows Normal, Depth, Albedo, and More

Jun 01, 2023
Figure 1 for StyleGAN knows Normal, Depth, Albedo, and More
Figure 2 for StyleGAN knows Normal, Depth, Albedo, and More
Figure 3 for StyleGAN knows Normal, Depth, Albedo, and More
Figure 4 for StyleGAN knows Normal, Depth, Albedo, and More
Viaarxiv icon

Make It So: Steering StyleGAN for Any Image Inversion and Editing

Add code
Apr 27, 2023
Figure 1 for Make It So: Steering StyleGAN for Any Image Inversion and Editing
Figure 2 for Make It So: Steering StyleGAN for Any Image Inversion and Editing
Figure 3 for Make It So: Steering StyleGAN for Any Image Inversion and Editing
Figure 4 for Make It So: Steering StyleGAN for Any Image Inversion and Editing
Viaarxiv icon