{"id":"b60bb022-231b-44a1-b189-3446ac1c4346","arxiv_id":"2502.05953","paper_version":1,"verdict":"REJECT","confidence":"MODERATE","novelty_score":2.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":2,"one_line_summary":"Mul2MAR combines ARToolKit markers, OpenGL rendering, and red-cyan anaglyph glasses to show virtual objects in apparent 3D on a mobile device, but gives no quantitative validation beyond the author's prior work.","lead":"This paper presents Mul2MAR, a marker-based augmented reality app that displays virtual furniture on a phone screen for viewing through inexpensive red-cyan 3D glasses. It claims to provide comfortable stereoscopic AR depth, but it offers only screenshots and leans on the author's own earlier 2014 paper for the core effect.","discovery_kind":"incremental","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The central comfort claim is internally contradicted by the paper's own Section 5 caveats and is backed by no user data; the claimed 3D perception is likewise asserted from screenshots and a self-citation rather than measured.","rationale":"The reader's REJECT verdict is appropriate. The paper's strongest claim has two parts, perception and comfort. The perception part is plausible for a standard anaglyph pipeline, so I do not object to the general idea; the single-marker stereo method follows from ARToolKit and OpenGL and is supported by the author's 2014 paper. However, the comfort part is not merely unmeasured; it is contradicted by the paper's own Section 5 limitations. The multi-marker contribution is also presented only through hand-tuned translations and scalings, exactly as the reader's weakest_assumption notes. My concrete test would settle whether the comfort/3D-perception claim survives, but absent that evidence the current manuscript is too thin to accept. Since my concern reinforces rather than changes the reader's rejection, the verdict stays REJECT/UNCHANGED.","tokens_in":7740,"tokens_out":7617,"duration_ms":80377,"concrete_test":"Conduct a preregistered within-subjects user study with at least 20 participants (including several who normally wear prescription glasses): each participant uses the app for 10 minutes while performing a depth-ordering task on the three furniture objects under red-cyan glasses, with pre/post Simulator Sickness Questionnaire and a discomfort VAS. If depth-order accuracy is not significantly above chance, or if mean SSQ/discomfort increases beyond a pre-registered threshold (or any participant reports persistent color afterimages), the claims of 3D perception and no uncomfortable situation are falsified.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The load-bearing condition for the abstract claim is that the app yields genuine stereoscopic depth 'without creating any uncomfortable situation for human eyes.' Section 5 explicitly concedes that prolonged red-cyan use may cause temporary colour impairment and that prescription-glasses wearers may not be able to use the glasses, which undercuts the no-discomfort part of the claim. No user study, discomfort metric, stereo-acuity test, or viewing-distance/IPD calibration is reported; the validation is a set of screenshots plus a self-citation [Güngör and Kurt 2014]. In Section 4, multi-object alignment is achieved by hand-editing code ('translating the objects on their positions of markers and scaling them'), so the multi-marker rendering is not shown to generalize. The stereo specification in Section 3 (shifting gl_para for left/right eyes) omits the projection parameters needed to control parallax and zero-parallax plane, so the comfortable-fusion condition is not demonstrated. Because the central claim depends on these unverified conditions, the paper's conclusion does not follow from its evidence.","agreement_with_reader":"partial"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper describes Mul2MAR, a mobile marker-based augmented reality application that uses ARToolKit for QR-style marker tracking and OpenGL for rendering. Left- and right-eye views are generated by shifting the gl_para matrix and are separated with glColorMask for viewing through red-cyan anaglyph glasses. The authors render several textured furniture objects with single and multiple markers and claim, in the abstract and conclusion, that the application provides satisfactory 3D AR perception without causing any uncomfortable situation for human eyes. Validation is presented as screenshots (Figures 4 and 5) and a citation to the authors' own prior work [Güngör and Kurt 2014].","tokens_in":7915,"tokens_out":6613,"duration_ms":59423,"significance":"If the perceptual claim were established, the application would be a useful low-cost demonstration that commodity mobile devices plus red-cyan glasses can produce a stereo AR experience. The paper is candid about a number of limitations in Section 5 and about the ad hoc nature of multi-marker placement in Section 4. However, the manuscript supplies no machine-checked proofs, reproducible code, or falsifiable measurements: there is no user study, no parallax or fusion-range measurement, no marker-detection statistics, no latency measurement, and no comfort/discomfort protocol. The contribution as written is therefore a qualitative demo note, and the abstract and conclusion make claims that go substantially beyond the evidence provided.","major_comments":[{"comment":"The load-bearing claim that the application provides “satisfactory 3D AR perception” and works “without creating any uncomfortable situation for human eyes” is not tested. The only evidence offered is the rendered screenshots and the sentence “When the scene is viewed with red-cyan glasses (see Figure 3), the visual perception of three-dimensional AR is provided [Güngör and Kurt 2014].” No stereo-acuity test, no user study, no comfort/discomfort metric, no measurement of viewing distance or interpupillary distance, and no parallax or fusion-range measurement is reported. Citation to the authors' own prior work does not validate this implementation, since the present application is a different system with different markers and scene configurations.","section":"Abstract and Section 3"},{"comment":"Section 5 explicitly concedes two downsides: “prolonged use of glasses may cause temporary colour impairment on each eye” and “people who use prescribed glasses may not be able to use 3D glasses easily.” These statements are in tension with the abstract's unconditional claim of operating “without creating any uncomfortable situation for human eyes.” At minimum, the central claim must be qualified to the tested viewing durations and user group; as written, the paper's own limitations contradict the headline claim.","section":"Section 5"},{"comment":"The multi-marker contribution is validated only through manual per-scene adjustments: “To solve this problem we made changes on the code, translating the objects on their positions of markers and scaling them.” No automatic registration, calibration procedure, or repeatability test is described, and no detection statistics are given despite the admission that “the system could not fully recognized some markers and objects were not displayed correctly by using these markers.” Consequently, the paper does not show that multiple virtual objects are correctly registered for arbitrary marker configurations, cameras, or users.","section":"Section 4"},{"comment":"The stereo camera description is incomplete in a way that matters for the comfort claim. The text says the camera is “shifted to the left and right eye points by modifying gl_para vector,” but it does not specify the interocular distance, the projection matrix, the near/far plane values, the focus-plane distance, or how the zero-parallax plane is established. These are exactly the parameters that control perceived depth and stereo discomfort, so the description provides no basis for assessing whether comfortable fusion with red-cyan glasses is achieved.","section":"Section 3"}],"minor_comments":[{"comment":"The caption refers to panels (b), (c), (e), (f), (h), and (k), but the figure labels jump from (h) to (k) without showing (i) or (j); please renumber the panels so that every referenced label exists.","section":"Figure 4 caption"},{"comment":"The paper does not report the mobile device model, screen size, camera resolution, ARToolKit version, or the exact source-code modifications used to produce Figures 4 and 5; these details are needed to reproduce or interpret the results.","section":"Reproducibility details"},{"comment":"The abstract uses “3D AR perception” as if it were a well-defined measurable outcome; please define the intended perceptual quality metric or replace this phrase with the observable that was actually recorded.","section":"Abstract"}],"recommendation":"reject","confidential_remarks":"For the editor: I recommend rejection. The central claims are unvalidated, and the paper's own Section 5 caveats undercut the “no discomfort” assertion. A revision would require a substantial new evaluation component—user studies, stereo parameter specifications, marker-detection statistics, and a repeatable multi-marker calibration procedure—rather than local fixes. I also note that the manuscript is formatted as an ACM TOG article but is more appropriate for a demo or short workshop paper; as submitted, the evidence supports only a demonstration, not the stated conclusions."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Quick take: this is a thin incremental extension of the author's own 2014 anaglyph AR paper. The only new bit is rendering multiple furniture objects on multiple QR-like markers, which is routine ARToolKit plus some hand-tuned translations and scalings. There is no user study, no parallax/latency measurement, and the abstract's comfort claim is undercut by the paper's own Section 5 caveats.\n\nWhat the paper does well: it is clearly written and short, and it is honest about the limitations of red-cyan glasses (temporary colour impairment, prescription glasses issues). The screenshots do show that the app displays several objects on markers, and the author explicitly credits his earlier paper for the core stereo technique rather than pretending it is brand new.\n\nSoft spots: the central claim, in the abstract and intro, is that the app provides 'satisfactory 3D AR perception without causing any uncomfortable situation for human eyes.' Section 5 then says prolonged use may cause temporary colour impairment and that people who wear prescribed glasses may not be able to use the 3D glasses easily. Those are uncomfortable situations. The abstract's blanket claim does not survive contact with the paper's own text. The validation is screenshots plus a self-citation to Güngör and Kurt 2014; no quantitative evidence for depth perception or comfort is reported. The stereo description says the camera is shifted by modifying gl_para, but no numbers are given for interocular distance, near/far planes, or the zero-parallax plane, so a reader cannot assess whether fusion would be comfortable or even correct. The multi-object rendering was achieved by hand-editing per-scene translations and scales, which means the demonstrated alignment does not generalize to new configurations.\n\nI agree with the reader's REJECT verdict. This is not a result for a journal like ACM TOG. It is a workshop or demo-level submission at best, and even then the unsupported comfort claim would have to be dropped or substantially qualified. No code, data, or precise parameters are provided, so the app as described is not reproducible. I would desk reject it.\n\nRecommendation: do not send this to peer review; desk reject. If the author wants to make a serious case, the next version needs a proper user study with stereo-acuity and comfort metrics, reported IPD and viewing distances, and a more cautious claim about discomfort.","headline":"A thin, self-cited extension of the author's 2014 anaglyph AR work; the multi-marker addition is routine, and the comfort claim is contradicted by the paper's own limitations section.","tokens_in":8482,"tokens_out":2085,"would_cite":false,"duration_ms":20719,"reading_group":"no","serious_thinker":"yes","would_accept_peer_review":false},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":[],"pacs":[],"model":"deepseek-v4-flash","headline":"A mobile app using printed markers and cheap red-cyan glasses can render several textured 3D objects in stereo, giving AR depth perception without discomfort, the paper claims.","keywords":["augmented reality","marker-based AR","anaglyph 3D","red-cyan glasses","multi-marker tracking","ARToolKit","OpenGL color masking","mobile AR"],"falsifier":"Take the app and a marker board that places the three QR-like markers in a layout not used in the paper, render a frame, and compare projected virtual-object positions against ground truth from a calibrated camera; if the objects drift from their markers, or if a user study measuring stereo acuity and comfort (for example, a forced-choice depth discrimination task) shows no depth benefit or reports discomfort, the central claims fail for general scenes.","tokens_in":7487,"feed_emoji":"🥽","tokens_out":3823,"duration_ms":37853,"temperature":0.7,"pith_summary":"The paper tries to establish that genuine stereo depth perception in augmented reality does not require expensive AR headsets: a regular mobile device, printable QR-like markers, and cheap red-cyan glasses suffice. It presents Mul2MAR, a marker-based AR application that renders several virtual objects—three furniture models—on multiple markers, with OpenGL producing separate left- and right-eye views that are color-masked and combined. The claim is that these combined views produce a convincing three-dimensional AR experience \"without creating any uncomfortable situation for human eyes.\" If correct, the practical consequence is that 3D AR becomes accessible at almost no hardware cost.","feed_headline":"Cheap red-cyan glasses make phone AR show depth","feed_subtitle":"Mul2MAR renders multiple textured objects in stereo on ARToolKit markers, no expensive headset required.","key_machinery":"The load-bearing mechanism is the modification of the OpenGL modelview matrix obtained from ARToolKit (arGetTransMat, argConvGlpara, glLoadMatrixd): the matrix is shifted laterally for the left and right eye positions before loading, and glColorMask(true, false, false, false) versus glColorMask(false, true, true, false) separates the two views into the red and cyan channels. QR-code-like markers are used instead of plain black-and-white patterns to improve detection. A second, less principled component is the manual translation and scaling of objects in code to align them with their markers.","core_discovery":"On the paper's own terms, the central discovery is that a complete stereo AR pipeline can be assembled from ARToolKit marker tracking, OpenGL rendering, and anaglyph color filtering: the device camera sees one marker scene, while the OpenGL camera is shifted to left- and right-eye viewpoints, and glColorMask selects red for one eye and cyan for the other. Multiple objects placed on several QR-like markers are then perceived in depth when viewed through red-cyan glasses, and the color fringing seen without the glasses disappears with them. The author reports that this was validated on single- and multi-object scenes using three furniture objects with assigned material properties and textures.","pith_inferences":["The paper does not include a user study measuring depth perception or comfort; \"without uncomfortable situation\" is an assertion, and a quantitative stereo-acuity test would be the natural next step.","Because registration relies on hand-tuned object translations per scene, an automated calibration procedure (for example, estimating marker-relative poses from the printed layout) would be needed before the multi-marker claim generalizes to arbitrary marker arrangements.","The approach inherits the known limitations of anaglyph stereo, such as color rivalry, so the claimed improvement in visual perception is likely strongest for low-chroma textured objects and less reliable for highly saturated or red-dominant scenes.","A direct comparison with head-mounted AR displays on the same depth-discrimination task would clarify whether the inexpensive anaglyph route provides practically useful depth or merely a novelty effect."],"forward_implications":["If the central claim holds, 3D AR on phones becomes nearly free: print a marker, buy red-cyan glasses, and run the app.","The same color-masking approach can be toggled on and off with a slight adjustment, so users can switch between normal and stereoscopic AR views.","Multiple objects can be displayed at once, not just single-marker demos, which extends the technique beyond trivial one-object examples.","Because the pipeline relies only on ARToolKit tracking and OpenGL rendering, it runs on ordinary mobile hardware without specialized depth sensors.","The results suggest that the stated future work—faster QR detection and deep-learning enhancements—could improve stability without changing the core stereoscopic mechanism."],"supporting_citations":[{"why":"Supplies the prior method of combining red-cyan glasses with mobile AR that this paper extends to multiple markers.","marker":"[Güngör and Kurt 2014]"},{"why":"Provides the marker detection and coordinate-transformation pipeline that the app relies on for registration.","marker":"[ARToolKit 2024]"},{"why":"Defines AR and the three attributes (combination, real-time interactivity, 3D registration) that frame the paper's validation.","marker":"[Azuma 1997]"}],"fun_headline_variants":["Multi-marker AR depth with cheap red-cyan glasses","Stereo AR on phone via markers and anaglyph glasses","No headset: phone AR gets depth from red-cyan glasses","Phone AR turns multi-marker scenes into 3D with anaglyph","Inexpensive glasses make phone AR images pop in 3D"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The multi-marker result rests on hand-tuned per-object translations and scalings described in Section 4; if those adjustments are specific to the tested furniture scenes, the claim that multiple objects render correctly does not extend to new marker layouts.","fun_headline_variants_meta":{"raw":{"variants":["Multi-marker AR depth with cheap red-cyan glasses","Stereo AR on phone via markers and anaglyph glasses","No headset: phone AR gets depth from red-cyan glasses","Phone AR turns multi-marker scenes into 3D with anaglyph","Inexpensive glasses make phone AR images pop in 3D"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000212,"raw_usage":{"total_tokens":1326,"prompt_tokens":762,"completion_tokens":564,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":378,"completion_tokens_details":{"reasoning_tokens":471}},"tokens_in":378,"tokens_out":564,"duration_ms":5749,"temperature":1.0,"reasoning_tokens":471,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-08T17:15:07.936325+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Take the app and a marker board that places the three QR-like markers in a layout not used in the paper, render a frame, and compare projected virtual-object positions against ground truth from a calibrated camera; if the objects drift from their markers, or if a user study measuring stereo acuity and comfort (for example, a forced-choice depth discrimination task) shows no depth benefit or reports discomfort, the central claims fail for general scenes.","supporting_citations":[],"review_version":1}