<?xml version="1.0" encoding="utf-8" standalone="yes"?><rss version="2.0" xmlns:atom="http://www.w3.org/2005/Atom"><channel><title>Monocular Depth Estimation on Jaehun's Blog</title><link>https://jaehun.me/tags/monocular-depth-estimation/</link><description>Recent content in Monocular Depth Estimation on Jaehun's Blog</description><generator>Hugo</generator><language>ko-kr</language><lastBuildDate>Sun, 13 Sep 2026 09:29:41 +0900</lastBuildDate><atom:link href="https://jaehun.me/tags/monocular-depth-estimation/index.xml" rel="self" type="application/rss+xml"/><item><title>Marigold V2: Revisiting Diffusion Transformers for Monocular Depth Estimation</title><link>https://jaehun.me/posts/marigold-v2-revisiting-diffusion-transformers-for-monocular-depth-estimation/</link><pubDate>Sun, 13 Sep 2026 00:00:00 +0900</pubDate><guid>https://jaehun.me/posts/marigold-v2-revisiting-diffusion-transformers-for-monocular-depth-estimation/</guid><description>&lt;p&gt;&lt;a&#10; href="https://arxiv.org/abs/2609.08084"target="_blank"&#10; class="inline-flex items-center gap-1"&#10; &gt;논문 링크&lt;svg class="h-3 w-3 flex-shrink-0" id="external-link" xmlns="http://www.w3.org/2000/svg" viewBox="0 0 24 24"&gt;&lt;path fill="none" stroke="currentColor" stroke-linecap="round" stroke-linejoin="round" stroke-width="2" d="M15 3h6v6m-11 5L21 3m-3 10v6a2 2 0 0 1-2 2H5a2 2 0 0 1-2-2V8a2 2 0 0 1 2-2h6"/&gt;&lt;/svg&gt;&#10; &lt;/a&gt;&lt;/p&gt;&#10;&lt;h2 id="marigold-v2-이미지-편집-확산-트랜스포머dit를-단안-깊이-추정기로-되살리기"&gt;Marigold V2: 이미지 편집 확산 트랜스포머(DiT)를 단안 깊이 추정기로 되살리기&lt;a href="#marigold-v2-%ec%9d%b4%eb%af%b8%ec%a7%80-%ed%8e%b8%ec%a7%91-%ed%99%95%ec%82%b0-%ed%8a%b8%eb%9e%9c%ec%8a%a4%ed%8f%ac%eb%a8%b8dit%eb%a5%bc-%eb%8b%a8%ec%95%88-%ea%b9%8a%ec%9d%b4-%ec%b6%94%ec%a0%95%ea%b8%b0%eb%a1%9c-%eb%90%98%ec%82%b4%eb%a6%ac%ea%b8%b0" class="heading-anchor" aria-label="이 섹션에 대한 링크"&gt;&lt;svg class="h-4 w-4" aria-hidden="true" xmlns="http://www.w3.org/2000/svg" viewBox="0 0 24 24"&gt;&lt;g fill="none" stroke="currentColor" stroke-linecap="round" stroke-linejoin="round" stroke-width="2"&gt;&lt;path d="M10 13a5 5 0 0 0 7.54.54l3-3a5 5 0 0 0-7.07-7.07l-1.72 1.71"/&gt;&lt;path d="M14 11a5 5 0 0 0-7.54-.54l-3 3a5 5 0 0 0 7.07 7.07l1.71-1.71"/&gt;&lt;/g&gt;&lt;/svg&gt;&lt;/a&gt;&lt;/h2&gt;&lt;h2 id="한-줄-요약-tldr"&gt;한 줄 요약 (TL;DR)&lt;a href="#%ed%95%9c-%ec%a4%84-%ec%9a%94%ec%95%bd-tldr" class="heading-anchor" aria-label="이 섹션에 대한 링크"&gt;&lt;svg class="h-4 w-4" aria-hidden="true" xmlns="http://www.w3.org/2000/svg" viewBox="0 0 24 24"&gt;&lt;g fill="none" stroke="currentColor" stroke-linecap="round" stroke-linejoin="round" stroke-width="2"&gt;&lt;path d="M10 13a5 5 0 0 0 7.54.54l3-3a5 5 0 0 0-7.07-7.07l-1.72 1.71"/&gt;&lt;path d="M14 11a5 5 0 0 0-7.54-.54l-3 3a5 5 0 0 0 7.07 7.07l1.71-1.71"/&gt;&lt;/g&gt;&lt;/svg&gt;&lt;/a&gt;&lt;/h2&gt;&lt;p&gt;Qwen-Image-Edit-2509라는 **이미지 편집용 확산 트랜스포머(DiT)**를 4-bit QLoRA로 파인튜닝해, &lt;strong&gt;단일 32GB GPU에서 약 1주일 만에&lt;/strong&gt; 단안 깊이 추정 SOTA를 달성한 연구다. 핵심은 두 가지 새로운 손실 — 깊이 GT에서 뽑은 의미 특징과 정렬하는 &lt;strong&gt;iREPA-depth&lt;/strong&gt;, 그리고 최적 수송(Sinkhorn) 기반의 &lt;strong&gt;SinkLoss&lt;/strong&gt; — 이며, 털·잎사귀·머리카락 같은 미세 디테일까지 살리면서 비행 픽셀(flying pixel) 아티팩트를 억제한다.&lt;/p&gt;</description></item></channel></rss>