Kazuyoshi Fushinobu | Searcharxiv

arXiv2026

PureCLIP-Depth: Prompt-Free and Decoder-Free Monocular Depth Estimation within CLIP Embedding Space

We propose PureCLIP-Depth, a completely prompt-free, decoder-free Monocular Depth Estimation (MDE) model that operates entirely within the Contrastive Language-Image Pre-training (CLIP) embedding space. Unlike recent models that rely heavily on geometric features, we explore a novel approach to MDE driven by conceptual information, performing computations directly within the conceptual CLIP space. The core of our method lies in learning a direct mapping from the RGB domain to the depth domain strictly inside this embedding space. Our approach achieves state-of-the-art performance among CLIP embedding-based models on both indoor and outdoor datasets. The code used in this research is available at: https://github.com/ryutaroLF/PureCLIP-Depth

cs.CV↗

arXiv2020

Selective laser ablation of metal thin films using ultrashort pulses

Selective thin-film removal is needed in many microfabrication processes such as 3-D patterning of optoelectronic devices and localized repairing of integrated circuits. Various wet or dry etching methods are available, but laser machining is a tool of green manufacturing as it can remove thin films by ablation without use of toxic chemicals. However, laser ablation causes thermal damage on neighboring patterns and underneath substrates, hindering its extensive use with high precision and integrity. Here, using ultrashort laser pulses of sub-picosecond duration, we demonstrate an ultrafast mechanism of laser ablation that leads to selective removal of a thin metal film with minimal damage on the substrate. The ultrafast laser ablation is accomplished with the insertion of a transition metal interlayer that offers high electron-phonon coupling to trigger vaporization in a picosecond timescale. This contained form of heat transfer permits lifting off the metal thin-film layer while blocking heat conduction to the substrate. Our ultrafast scheme of selective thin film removal is analytically validated using a two-temperature model of heat transfer between electrons and phonons in material. Further, experimental verification is made using 0.2 ps laser pulses by micropatterning metal films for various applications.

physics.app-ph↗